Computer Incident Response Center Luxembourg logo
Organization

Computer Incident Response Center Luxembourg

Governmental, Luxembourg

GNA-1

CIRCL is the CERT (Computer Emergency Response Team/Computer Security Incident Response Team) for the private sector, communes and non-governmental entities in Luxembourg.


Website
https://www.circl.lu
Email
info@circl.lu
Vendor name
circl
GNA identifier
GNA-1 GCVE registry Recent publications

Products

9 products registered by Computer Incident Response Center Luxembourg

Open Source Threat Intelligence Platform & Open Standards For Threat Information Sharing

Client API to query any Passive DNS implementation following the Passive DNS - Common Output Format.

An open source platform to support analysts to organise their case and tasks.

AIL Framework

The Vulnerability-Lookup software is a powerful sharing platform to assist security teams, researchers, and system administrators in identifying and tracking vulnerabilities related to specific vendors and products. By leveraging known vulnerability databases, sightings, and configurations, the software provides...

Analyse a forensic target (such as a directory) to find and report files found and not found from CIRCL hashlookup public service or the Bloom filter from CIRCL hashlookup. This tool can help a digital forensic investigator to know the context, origin of specific files during a digital forensic investigation.

Lookyloo is a web interface that captures a webpage and then displays a tree of the domains, that call each other.

Convert Cyber Threat Intelligence data between MISP and STIX formats. Open-source REST API maintained by the MISP Project community.

Rulezet is an open-source, community-driven platform for sharing cybersecurity detection rules. This policy covers the official instance at rulezet.org. Anyone can also self-host Rulezet from its source code — if you are using a self-hosted instance, this default policy may have been customized by that instance's...

Recent vulnerabilities

CVE-2026-133848 (GCVE-0-2026-133848)

Vulnerability from cvelistv5 – Published: 2026-11-20 14:39 – Updated: 2026-11-20 18:31
VLAI
Title
flowintel: Organization Admin Can Delete Platform Admin Accounts Due to Missing Privilege Hierarchy Check
Summary
flowintel contains an authorization flaw in its user deletion functionality. Organization-scoped administrators (org admins) are able to delete platform-level administrator accounts through both the browser-based admin interface and the REST API. The existing authorization logic verified that an org admin could only target users within their own organization, but it did not enforce the privilege hierarchy between organization admins and platform admins. As a result, an authenticated org admin could issue a deletion request against any platform-admin account belonging to the same organization, causing the target account, its sessions, and related records to be permanently removed. Impact: - Loss of platform-level administrative oversight - Potential disruption of platform-wide management functions - Removal of higher-privilege accounts that should be protected from lower-privilege roles Affected versions: flowintel prior to 4.1.0
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 14:19 UTC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:31 UTC
CWE
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 14:19
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb 05b89a0d2175…
Confidence
medium
Commit Subject Patch SHA-256
e3d0b2ab153d fix: [security] Prevent organization admins from deleting 05b89a0d2175…
Fix summary

The fix introduces an explicit privilege-hierarchy check in both the browser and API deletion routes. Before any session invalidation or account deletion is performed, the system now verifies whether the target user holds platform-admin capability. If the requester is a pure organization admin and the target is a platform admin, the request is rejected with a 403 response, preserving the intended authorization boundary while leaving legitimate deletion workflows unaffected.

Patch summary

Two lines added in app/admin/admin.py (delete_user function) and two lines added in app/admin/admin_api.py (API GET handler). In each location, after the existing same-organization check for pure org admins, a new condition checks user.is_admin() / user_to_delete.is_admin(). If true, a 403 JSON response is returned with an appropriate error message, preventing the deletion from proceeding. No other logic was modified.

CVSS rationale

AV:N - the vulnerability is exploitable over the network via the web UI or REST API. AC:L - exploitation requires only a standard authenticated request targeting a platform-admin account; no race conditions or special timing. AT:N - no attack-target manipulation is needed. PR:L - the attacker must hold a valid organization-admin account. UI:N - no victim interaction is required; the org admin acts on their own behalf. VC:N - no confidential data is exposed. VI:N - the application's own data integrity is not corrupted; the deletion is a legitimate operation performed by an authenticated user. VA:H - platform-admin accounts and their sessions are permanently deleted, significantly degrading the availability of platform administration. SC:N - no cross-component confidentiality impact. SI:N - no cross-component integrity impact. SA:H - the broader platform administration and oversight functions lose availability when platform-admin accounts are removed.

Weakness rationale
  • CWE-285 The authorization logic for the user-deletion endpoint verified organizational scope but failed to enforce the privilege hierarchy between organization admins and platform admins. The missing check allowed a lower-privilege role to perform an action (deleting a higher-privilege account) that should have been restricted, constituting an improper authorization decision.
Attack pattern rationale
  • CAPEC-126 An authenticated organization admin leverages a missing authorization boundary to perform an action (deleting a platform-admin account) that is reserved for a higher-privilege role. This is a vertical privilege escalation: the actor gains the effective capability of a platform admin by removing platform-admin accounts. The mapping is the closest standard CAPEC; the attack does not involve traditional credential theft or token manipulation but rather an authorization logic gap that permits a lower-privilege actor to act at a higher privilege level.
Assumptions to verify
  • The affected version range is inferred from the explicit context (fixed in 4.1.0); the exact earliest affected version is not stated in the patch and is recorded as unspecified.
  • The CAPEC-126 mapping is the closest standard pattern; the vulnerability is more precisely an authorization-logic gap than a classic privilege-escalation exploit, but no more specific CAPEC exists for missing privilege-hierarchy checks in multi-tenant admin systems.
  • CVSS VA:H and SA:H assume that deleting platform-admin accounts materially degrades the availability of platform-level administration; if the deployment has redundant platform admins, the practical impact may be lower.
  • The patch metadata lists affected_version as '0', which is treated as a placeholder; the actual lower bound of affected versions is unknown.
  • No evidence of active exploitation or public PoC was found in the patch or metadata.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 7 11 medium 5
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-17 08:09 UTC Fix developed Corrective change authored (e3d0b2ab153d74df72a1f0c417dc2eef74554bcb): fix: [security] Prevent organization admins from deleting https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb
  2. 2026-11-20 14:18 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133848",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:31:26.526789Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:31:42.592Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/admin/admin.py",
            "app/admin/admin_api.py"
          ],
          "product": "flowintel",
          "programFiles": [
            "app/admin/admin.py",
            "app/admin/admin_api.py"
          ],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "flowintel contains an authorization flaw in its user deletion functionality. Organization-scoped administrators (org admins) are able to delete platform-level administrator accounts through both the browser-based admin interface and the REST API.\n\nThe existing authorization logic verified that an org admin could only target users within their own organization, but it did not enforce the privilege hierarchy between organization admins and platform admins. As a result, an authenticated org admin could issue a deletion request against any platform-admin account belonging to the same organization, causing the target account, its sessions, and related records to be permanently removed.\n\nImpact:\n- Loss of platform-level administrative oversight\n- Potential disruption of platform-wide management functions\n- Removal of higher-privilege accounts that should be protected from lower-privilege roles\n\nAffected versions: flowintel prior to 4.1.0"
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-126",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-126 Privilege Escalation"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 8.3,
            "baseSeverity": "HIGH",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "HIGH",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:H",
            "version": "4.0",
            "vulnAvailabilityImpact": "HIGH",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "NONE",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T14:19:16Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-285",
              "description": "CWE-285 Improper Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T14:39:11.588Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The fix introduces an explicit privilege-hierarchy check in both the browser and API deletion routes. Before any session invalidation or account deletion is performed, the system now verifies whether the target user holds platform-admin capability. If the requester is a pure organization admin and the target is a platform admin, the request is rejected with a 403 response, preserving the intended authorization boundary while leaving legitimate deletion workflows unaffected."
        }
      ],
      "title": "flowintel: Organization Admin Can Delete Platform Admin Accounts Due to Missing Privilege Hierarchy Check",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected version range is inferred from the explicit context (fixed in 4.1.0); the exact earliest affected version is not stated in the patch and is recorded as unspecified.",
                  "The CAPEC-126 mapping is the closest standard pattern; the vulnerability is more precisely an authorization-logic gap than a classic privilege-escalation exploit, but no more specific CAPEC exists for missing privilege-hierarchy checks in multi-tenant admin systems.",
                  "CVSS VA:H and SA:H assume that deleting platform-admin accounts materially degrades the availability of platform-level administration; if the deployment has redundant platform admins, the practical impact may be lower.",
                  "The patch metadata lists affected_version as \u00270\u0027, which is treated as a placeholder; the actual lower bound of affected versions is unknown.",
                  "No evidence of active exploitation or public PoC was found in the patch or metadata."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-126",
                    "rationale": "An authenticated organization admin leverages a missing authorization boundary to perform an action (deleting a platform-admin account) that is reserved for a higher-privilege role. This is a vertical privilege escalation: the actor gains the effective capability of a platform admin by removing platform-admin accounts. The mapping is the closest standard CAPEC; the attack does not involve traditional credential theft or token manipulation but rather an authorization logic gap that permits a lower-privilege actor to act at a higher privilege level."
                  }
                ],
                "commit": "e3d0b2ab153d74df72a1f0c417dc2eef74554bcb",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N - the vulnerability is exploitable over the network via the web UI or REST API. AC:L - exploitation requires only a standard authenticated request targeting a platform-admin account; no race conditions or special timing. AT:N - no attack-target manipulation is needed. PR:L - the attacker must hold a valid organization-admin account. UI:N - no victim interaction is required; the org admin acts on their own behalf. VC:N - no confidential data is exposed. VI:N - the application\u0027s own data integrity is not corrupted; the deletion is a legitimate operation performed by an authenticated user. VA:H - platform-admin accounts and their sessions are permanently deleted, significantly degrading the availability of platform administration. SC:N - no cross-component confidentiality impact. SI:N - no cross-component integrity impact. SA:H - the broader platform administration and oversight functions lose availability when platform-admin accounts are removed.",
                "fixSummary": "The fix introduces an explicit privilege-hierarchy check in both the browser and API deletion routes. Before any session invalidation or account deletion is performed, the system now verifies whether the target user holds platform-admin capability. If the requester is a pure organization admin and the target is a platform admin, the request is rejected with a 403 response, preserving the intended authorization boundary while leaving legitimate deletion workflows unaffected.",
                "generatedAt": "2026-11-20T14:19:16.357167Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 5,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 7
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "05b89a0d21756a6aeff8a9274a8c1993f718a0d5253b3aeabfe7965dfe04c470",
                "patchSummary": "Two lines added in app/admin/admin.py (delete_user function) and two lines added in app/admin/admin_api.py (API GET handler). In each location, after the existing same-organization check for pure org admins, a new condition checks user.is_admin() / user_to_delete.is_admin(). If true, a 403 JSON response is returned with an appropriate error message, preventing the deletion from proceeding. No other logic was modified.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "e3d0b2ab153d74df72a1f0c417dc2eef74554bcb",
                    "date": "Tue, 17 Nov 2026 09:09:12 +0100",
                    "patchSha256": "05b89a0d21756a6aeff8a9274a8c1993f718a0d5253b3aeabfe7965dfe04c470",
                    "retrievedAt": "2026-11-20T14:18:48.516621Z",
                    "source": "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb",
                    "subject": "fix: [security] Prevent organization admins from deleting"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T14:19:16Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Prevent organization admins from deleting",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-285",
                    "rationale": "The authorization logic for the user-deletion endpoint verified organizational scope but failed to enforce the privilege hierarchy between organization admins and platform admins. The missing check allowed a lower-privilege role to perform an action (deleting a higher-privilege account) that should have been restricted, constituting an improper authorization decision."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (e3d0b2ab153d74df72a1f0c417dc2eef74554bcb): fix: [security] Prevent organization admins from deleting",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb"
                    ],
                    "timestamp": "2026-11-17T08:09:12Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb"
                    ],
                    "timestamp": "2026-11-20T14:18:48.516621Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20416"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133848",
    "datePublished": "2026-11-20T14:39:11.588Z",
    "dateReserved": "2026-11-20T14:39:06.507Z",
    "dateUpdated": "2026-11-20T18:31:42.592Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133667 (GCVE-0-2026-133667)

Vulnerability from cvelistv5 – Published: 2026-11-20 10:58 – Updated: 2026-11-20 18:26
VLAI
Title
flowintel DOM-based XSS via Timeline Autolinking, Search Highlighting, and Content-Type Confusion
Summary
flowintel contains multiple related flaws that allow an authenticated attacker to inject arbitrary HTML or script into pages viewed by other users. - Timeline rendering: the timeline library's autolinking feature rewrote already-sanitised case notes and imported MISP attribute values into generated HTML link elements. Crafted text containing quote characters could break out of the generated attribute context, altering the surrounding markup and enabling script execution in the victim's browser. - Search highlighting: the highlight injection logic used a regular expression to locate text between angle brackets and spliced <mark> elements into the resulting string. Because the pattern could match inside HTML attribute values, the rewrite could corrupt or inject content into attributes of already-sanitised markup, again leading to DOM-based cross-site scripting. - Content-Type confusion: three analyser selection endpoints returned session-stored JSON strings with a text/html content type. A browser receiving such a response will parse and render the body as HTML, so any HTML or script embedded in the stored value is executed in the victim's context. Preconditions: the attacker must be an authenticated user with the ability to create or import case content (notes, MISP attributes) containing specially crafted quote or angle-bracket sequences. The victim must subsequently open the affected case timeline, run a search, or request one of the selection endpoints. Impact: successful exploitation yields DOM-based cross-site scripting in the victim's browser session, potentially allowing disclosure of case data, session manipulation, or execution of arbitrary actions under the victim's identity. Affected versions: flowintel prior to 4.1.0.
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 10:50 UTC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:26 UTC
CWE
  • CWE-79 - Improper Neutralization of Input in Web Page ('Cross-site Scripting')
  • CWE-799 - Improper Control of Resources Identifiers
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 10:50
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b.patch 5352fc62c6a6…
Confidence
medium
Commit Subject Patch SHA-256
5f745040d79d fix: [security] Timeline and search rendering preserve text 5352fc62c6a6…
Fix summary

The vulnerability is remediated by eliminating the three injection paths: (1) autolinking is disabled on every timeline event so the library no longer rewrites sanitised HTML; (2) search and auto-highlight marks are now created by walking the DOM text nodes and setting textContent, which cannot alter existing attributes or split character entities; (3) the three selection endpoints now decode the stored JSON once, respond with application/json and X-Content-Type-Options: nosniff, and reject malformed or excessively nested payloads with a 400 JSON error, preventing the browser from interpreting the body as HTML.

Patch summary

In case_history.js and timeline_view.js, an autolink: false property is added to each TimelineJS event object to suppress linkification. In smart-render.js, the previous regex-based string replacement (html.replace(/>([^<]*)</g, ...)) is replaced with a DOM TreeWalker that iterates text nodes, builds <mark> elements via document.createElement and textContent, and splices them into a DocumentFragment, preserving all existing attributes and entities. A new count_html_matches helper counts matches across text nodes. The plain-text fallback branches that performed raw regex replacement on escaped strings are removed. In misp_modules.py, a new _selected_json_response helper decodes the session value with json.loads, wraps it in jsonify, sets X-Content-Type-Options: nosniff, and returns a 400 JSON error on TypeError, ValueError, or RecursionError. The three get_*_selected endpoints are updated to call this helper instead of returning the raw session string.

CVSS rationale

AV:N – the application is a web service reachable over the network. AC:L – crafting quote-bearing text in case notes or MISP attributes is straightforward; no race or special timing is needed. AT:N – no active attack or prior compromise is required beyond normal authenticated use. PR:L – the attacker needs an authenticated account to create or import the malicious content (endpoints are login_required). UI:A – the victim must actively open the case timeline, run a search, or request a selection endpoint for the injected markup to execute. VC:L – XSS can expose case data, session tokens, or other sensitive information visible in the browser. VI:L – the attacker can modify displayed content or trigger actions in the victim's session. VA:N – no evidence of denial-of-service impact beyond the nested-JSON guard. SC:N – no cross-component or cross-origin impact is demonstrated. SI:L – other authenticated users viewing the same case could be affected by the stored malicious content. SA:N – no elevation of security authority is indicated.

Weakness rationale
  • CWE-79 All three sub-issues (autolinking rewriting sanitised HTML, regex-based highlight injection corrupting attributes, and serving JSON as text/html) result in attacker-controlled content being interpreted as HTML/JavaScript in the victim's browser, which is the defining characteristic of DOM-based XSS.
  • CWE-799 The content-type confusion (serving application/json data as text/html) is a secondary weakness that amplifies the XSS by causing the browser to parse the response body as HTML. This is a narrower, supporting CWE.
Attack pattern rationale
  • CAPEC-6 The primary exploitation path is client-side: attacker-controlled text (case notes, MISP values, search terms) is processed by JavaScript in the browser (TimelineJS autolinking, regex-based highlight injection) and the resulting DOM manipulation injects executable markup. The content-type confusion endpoint is a secondary vector but the dominant pattern is DOM-based script injection. CAPEC-6 is the closest standard attack pattern; the mapping is confident because the patch explicitly addresses client-side HTML rewriting of user-supplied content.
  • CAPEC-1 The three selection endpoints that returned session JSON as text/html could be leveraged as a reflected XSS vector if the session content is influenced by a prior request parameter. This is a secondary, less certain mapping; CAPEC-6 remains the primary pattern. Included because the content-type confusion path is a distinct mechanism from pure DOM manipulation.
Assumptions to verify
  • The affected_version field in the metadata is recorded as '0', which is treated as a placeholder meaning all versions prior to 4.1.0 are affected; no earlier version boundary is specified.
  • The CVSS PR:L assumes the attacker needs an authenticated account to create the malicious case content; if unauthenticated import paths exist, PR could be N.
  • The CVSS UI:A assumes the victim must navigate to the affected page; if the malicious content is auto-loaded in a dashboard or notification, UI could be N.
  • CAPEC-1 (Reflected XSS) is included as a secondary mapping for the content-type confusion path; the primary and more certain mapping is CAPEC-6 (DOM Based XSS).
  • The nested-JSON rejection (RecursionError guard) suggests a potential DoS vector via deeply nested JSON, but no separate DoS impact is scored because the primary vulnerability is XSS.
  • No public exploit or PoC is referenced; exploitation status is assumed to be 'none'.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 6 11 medium 6
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-10-30 11:49 UTC Fix developed Corrective change authored (5f745040d79d2f08d4a06512a91313c5a4eb346b): fix: [security] Timeline and search rendering preserve text https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b.patch
  2. 2026-11-20 10:49 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133667",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:26:26.477604Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:26:39.347Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/analyzer/misp_modules.py",
            "app/static/js/case/case_history.js",
            "app/static/js/case/timeline_view.js",
            "app/static/js/components/smart-render.js"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "flowintel contains multiple related flaws that allow an authenticated attacker to inject arbitrary HTML or script into pages viewed by other users.\n\n- Timeline rendering: the timeline library\u0027s autolinking feature rewrote already-sanitised case notes and imported MISP attribute values into generated HTML link elements. Crafted text containing quote characters could break out of the generated attribute context, altering the surrounding markup and enabling script execution in the victim\u0027s browser.\n\n- Search highlighting: the highlight injection logic used a regular expression to locate text between angle brackets and spliced \u003cmark\u003e elements into the resulting string. Because the pattern could match inside HTML attribute values, the rewrite could corrupt or inject content into attributes of already-sanitised markup, again leading to DOM-based cross-site scripting.\n\n- Content-Type confusion: three analyser selection endpoints returned session-stored JSON strings with a text/html content type. A browser receiving such a response will parse and render the body as HTML, so any HTML or script embedded in the stored value is executed in the victim\u0027s context.\n\nPreconditions: the attacker must be an authenticated user with the ability to create or import case content (notes, MISP attributes) containing specially crafted quote or angle-bracket sequences. The victim must subsequently open the affected case timeline, run a search, or request one of the selection endpoints.\n\nImpact: successful exploitation yields DOM-based cross-site scripting in the victim\u0027s browser session, potentially allowing disclosure of case data, session manipulation, or execution of arbitrary actions under the victim\u0027s identity.\n\nAffected versions: flowintel prior to 4.1.0."
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-6",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-6 DOM Based XSS"
            }
          ]
        },
        {
          "capecId": "CAPEC-1",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-1 XSS - Reflected"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 4.8,
            "baseSeverity": "MEDIUM",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "LOW",
            "userInteraction": "ACTIVE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:A/VC:L/VI:L/VA:N/SC:N/SI:L/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "LOW",
            "vulnIntegrityImpact": "LOW",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T10:50:21Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-79",
              "description": "CWE-79 Improper Neutralization of Input in Web Page (\u0027Cross-site Scripting\u0027)",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-799",
              "description": "CWE-799 Improper Control of Resources Identifiers",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T10:58:27.052Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The vulnerability is remediated by eliminating the three injection paths: (1) autolinking is disabled on every timeline event so the library no longer rewrites sanitised HTML; (2) search and auto-highlight marks are now created by walking the DOM text nodes and setting textContent, which cannot alter existing attributes or split character entities; (3) the three selection endpoints now decode the stored JSON once, respond with application/json and X-Content-Type-Options: nosniff, and reject malformed or excessively nested payloads with a 400 JSON error, preventing the browser from interpreting the body as HTML."
        }
      ],
      "title": "flowintel DOM-based XSS via Timeline Autolinking, Search Highlighting, and Content-Type Confusion",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected_version field in the metadata is recorded as \u00270\u0027, which is treated as a placeholder meaning all versions prior to 4.1.0 are affected; no earlier version boundary is specified.",
                  "The CVSS PR:L assumes the attacker needs an authenticated account to create the malicious case content; if unauthenticated import paths exist, PR could be N.",
                  "The CVSS UI:A assumes the victim must navigate to the affected page; if the malicious content is auto-loaded in a dashboard or notification, UI could be N.",
                  "CAPEC-1 (Reflected XSS) is included as a secondary mapping for the content-type confusion path; the primary and more certain mapping is CAPEC-6 (DOM Based XSS).",
                  "The nested-JSON rejection (RecursionError guard) suggests a potential DoS vector via deeply nested JSON, but no separate DoS impact is scored because the primary vulnerability is XSS.",
                  "No public exploit or PoC is referenced; exploitation status is assumed to be \u0027none\u0027."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-6",
                    "rationale": "The primary exploitation path is client-side: attacker-controlled text (case notes, MISP values, search terms) is processed by JavaScript in the browser (TimelineJS autolinking, regex-based highlight injection) and the resulting DOM manipulation injects executable markup. The content-type confusion endpoint is a secondary vector but the dominant pattern is DOM-based script injection. CAPEC-6 is the closest standard attack pattern; the mapping is confident because the patch explicitly addresses client-side HTML rewriting of user-supplied content."
                  },
                  {
                    "capecId": "CAPEC-1",
                    "rationale": "The three selection endpoints that returned session JSON as text/html could be leveraged as a reflected XSS vector if the session content is influenced by a prior request parameter. This is a secondary, less certain mapping; CAPEC-6 remains the primary pattern. Included because the content-type confusion path is a distinct mechanism from pure DOM manipulation."
                  }
                ],
                "commit": "5f745040d79d2f08d4a06512a91313c5a4eb346b",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N \u2013 the application is a web service reachable over the network. AC:L \u2013 crafting quote-bearing text in case notes or MISP attributes is straightforward; no race or special timing is needed. AT:N \u2013 no active attack or prior compromise is required beyond normal authenticated use. PR:L \u2013 the attacker needs an authenticated account to create or import the malicious content (endpoints are login_required). UI:A \u2013 the victim must actively open the case timeline, run a search, or request a selection endpoint for the injected markup to execute. VC:L \u2013 XSS can expose case data, session tokens, or other sensitive information visible in the browser. VI:L \u2013 the attacker can modify displayed content or trigger actions in the victim\u0027s session. VA:N \u2013 no evidence of denial-of-service impact beyond the nested-JSON guard. SC:N \u2013 no cross-component or cross-origin impact is demonstrated. SI:L \u2013 other authenticated users viewing the same case could be affected by the stored malicious content. SA:N \u2013 no elevation of security authority is indicated.",
                "fixSummary": "The vulnerability is remediated by eliminating the three injection paths: (1) autolinking is disabled on every timeline event so the library no longer rewrites sanitised HTML; (2) search and auto-highlight marks are now created by walking the DOM text nodes and setting textContent, which cannot alter existing attributes or split character entities; (3) the three selection endpoints now decode the stored JSON once, respond with application/json and X-Content-Type-Options: nosniff, and reject malformed or excessively nested payloads with a 400 JSON error, preventing the browser from interpreting the body as HTML.",
                "generatedAt": "2026-11-20T10:50:21.529700Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 6,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 6
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "5352fc62c6a6482c29f863cbf7a788fdb94c38d87b96d115294e35c697eeefc6",
                "patchSummary": "In case_history.js and timeline_view.js, an autolink: false property is added to each TimelineJS event object to suppress linkification. In smart-render.js, the previous regex-based string replacement (html.replace(/\u003e([^\u003c]*)\u003c/g, ...)) is replaced with a DOM TreeWalker that iterates text nodes, builds \u003cmark\u003e elements via document.createElement and textContent, and splices them into a DocumentFragment, preserving all existing attributes and entities. A new count_html_matches helper counts matches across text nodes. The plain-text fallback branches that performed raw regex replacement on escaped strings are removed. In misp_modules.py, a new _selected_json_response helper decodes the session value with json.loads, wraps it in jsonify, sets X-Content-Type-Options: nosniff, and returns a 400 JSON error on TypeError, ValueError, or RecursionError. The three get_*_selected endpoints are updated to call this helper instead of returning the raw session string.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "5f745040d79d2f08d4a06512a91313c5a4eb346b",
                    "date": "Fri, 30 Oct 2026 12:49:38 +0100",
                    "patchSha256": "5352fc62c6a6482c29f863cbf7a788fdb94c38d87b96d115294e35c697eeefc6",
                    "retrievedAt": "2026-11-20T10:49:45.815552Z",
                    "source": "https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b.patch",
                    "subject": "fix: [security] Timeline and search rendering preserve text"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T10:50:21Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Timeline and search rendering preserve text",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-79",
                    "rationale": "All three sub-issues (autolinking rewriting sanitised HTML, regex-based highlight injection corrupting attributes, and serving JSON as text/html) result in attacker-controlled content being interpreted as HTML/JavaScript in the victim\u0027s browser, which is the defining characteristic of DOM-based XSS."
                  },
                  {
                    "cweId": "CWE-799",
                    "rationale": "The content-type confusion (serving application/json data as text/html) is a secondary weakness that amplifies the XSS by causing the browser to parse the response body as HTML. This is a narrower, supporting CWE."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (5f745040d79d2f08d4a06512a91313c5a4eb346b): fix: [security] Timeline and search rendering preserve text",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b.patch"
                    ],
                    "timestamp": "2026-10-30T11:49:38Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b.patch"
                    ],
                    "timestamp": "2026-11-20T10:49:45.815552Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20406"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133667",
    "datePublished": "2026-11-20T10:58:27.052Z",
    "dateReserved": "2026-11-20T10:58:25.195Z",
    "dateUpdated": "2026-11-20T18:26:39.347Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133665 (GCVE-0-2026-133665)

Vulnerability from cvelistv5 – Published: 2026-11-20 10:58 – Updated: 2026-11-20 18:26
VLAI
Title
flowintel Information Disclosure of Webhook Secrets and Logs to Non-Admin Users
Summary
flowintel contains an information disclosure vulnerability that allows authenticated non-administrative users to access sensitive configuration data and logs intended for administrators only. The notification page and alert API responses exposed sensitive fields, including webhook URLs, signing secrets, IMAP credentials, and delivery logs, in the initial data or JSON responses. Although the user interface hid the configuration form from non-admin users, the underlying data remained accessible via the page source or API endpoints. This exposure included historical webhook URLs that may contain active bearer tokens. An attacker with a valid non-admin account could retrieve these credentials and logs, potentially leading to unauthorized access to notification services, impersonation of the application in webhook deliveries, or further lateral movement within the infrastructure. Affected versions: flowintel < 4.1.0
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 10:47 UTC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:25 UTC
CWE
  • CWE-200 - Exposure of Sensitive Information to an Unauthorized Actor
  • CWE-862 - Missing Authorization
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 10:47
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66.patch ee51a7b891c2…
Confidence
high
Commit Subject Patch SHA-256
1445157891ad fix: [security] Alert secrets and webhook logs stay ee51a7b891c2…
Fix summary

The vulnerability is remediated by enforcing role-based access control on data serialization and template rendering. Non-admin users are now restricted to receiving only public enable flags and do not have access to webhook URLs, secrets, or delivery logs. The shared alert serializer explicitly removes sensitive fields for non-admin users, ensuring that both the notification page and the alerts API preserve administrative diagnostics while protecting sensitive data from unauthorized access.

Patch summary

The patch modifies app/alerts/alerts_core.py and app/notification/notification.py. In alerts_core.py, the alerts_to_json function now checks if the user is an admin and removes the webhook_url field from the JSON response for non-admin users. In notification.py, the index route now checks the current user's admin status. For non-admins, it passes a restricted configuration object containing only public enable flags (CASE_CREATE_ALERT_ENABLED, WEBHOOK_ENABLED) instead of the full ConfigModule, and it passes an empty list for alert_logs instead of the actual log data. This ensures sensitive configuration and logs are not embedded in the template context for unauthorized users.

CVSS rationale

Attack Vector is Network (AV:N) as the application is web-based. Attack Complexity is Low (AC:L) as it requires only standard HTTP requests. Attack Requirements are None (AT:N). Privileges Required are Low (PR:L) as any authenticated non-admin user can exploit this. User Interaction is None (UI:N). Vulnerable Component Impact is None (VC:N) as the server itself is not compromised. Vulnerable Component Integrity is High (VI:H) because sensitive secrets and logs are exposed, which can lead to further compromise. Vulnerable Component Availability is None (VA:N). Subsequent Component Impact is None (SC:N, SI:N, SA:N) as the direct impact is limited to information disclosure within the vulnerable component, though downstream impacts are possible but not directly measured by the base vector.

Weakness rationale
  • CWE-200 The application exposed sensitive configuration data (webhook URLs, secrets, logs) to authenticated users who did not have the necessary privileges to view them.
  • CWE-862 The application failed to enforce authorization checks when serializing alert data and rendering the notification page, allowing non-admin users to access admin-only data.
Attack pattern rationale
  • CAPEC-126 The attacker leveraged their authenticated non-admin role to access data (webhook secrets, logs) that was not properly restricted to the administrator role, despite the UI hiding the configuration form.
Assumptions to verify
  • The affected version is assumed to be all versions prior to 4.1.0 based on the fixed_version metadata.
  • The CAPEC mapping is based on the general pattern of accessing restricted functionality, as the specific attack vector is a standard HTTP request to an existing endpoint.
  • The CVSS vector assumes that the exposure of secrets leads to a high impact on integrity of the vulnerable component, as these secrets can be used to impersonate the application or access other services.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 10 11 high 3
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-10-30 13:18 UTC Fix developed Corrective change authored (1445157891add7c04a63e0014cd96bf9802cac66): fix: [security] Alert secrets and webhook logs stay https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66.patch
  2. 2026-11-20 10:47 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133665",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:25:43.144657Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:26:03.887Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/alerts/alerts_core.py",
            "app/notification/notification.py"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "flowintel contains an information disclosure vulnerability that allows authenticated non-administrative users to access sensitive configuration data and logs intended for administrators only.\n\nThe notification page and alert API responses exposed sensitive fields, including webhook URLs, signing secrets, IMAP credentials, and delivery logs, in the initial data or JSON responses. Although the user interface hid the configuration form from non-admin users, the underlying data remained accessible via the page source or API endpoints. This exposure included historical webhook URLs that may contain active bearer tokens.\n\nAn attacker with a valid non-admin account could retrieve these credentials and logs, potentially leading to unauthorized access to notification services, impersonation of the application in webhook deliveries, or further lateral movement within the infrastructure.\n\nAffected versions: flowintel \u003c 4.1.0"
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-126",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-126 Accessing Functionality Not Properly Restricted by User Role"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 7.1,
            "baseSeverity": "HIGH",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:H/VA:N/SC:N/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "HIGH",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T10:47:41Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-200",
              "description": "CWE-200 Exposure of Sensitive Information to an Unauthorized Actor",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-862",
              "description": "CWE-862 Missing Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T10:58:14.508Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The vulnerability is remediated by enforcing role-based access control on data serialization and template rendering. Non-admin users are now restricted to receiving only public enable flags and do not have access to webhook URLs, secrets, or delivery logs. The shared alert serializer explicitly removes sensitive fields for non-admin users, ensuring that both the notification page and the alerts API preserve administrative diagnostics while protecting sensitive data from unauthorized access."
        }
      ],
      "title": "flowintel Information Disclosure of Webhook Secrets and Logs to Non-Admin Users",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected version is assumed to be all versions prior to 4.1.0 based on the fixed_version metadata.",
                  "The CAPEC mapping is based on the general pattern of accessing restricted functionality, as the specific attack vector is a standard HTTP request to an existing endpoint.",
                  "The CVSS vector assumes that the exposure of secrets leads to a high impact on integrity of the vulnerable component, as these secrets can be used to impersonate the application or access other services."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-126",
                    "rationale": "The attacker leveraged their authenticated non-admin role to access data (webhook secrets, logs) that was not properly restricted to the administrator role, despite the UI hiding the configuration form."
                  }
                ],
                "commit": "1445157891add7c04a63e0014cd96bf9802cac66",
                "confidence": "high",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "Attack Vector is Network (AV:N) as the application is web-based. Attack Complexity is Low (AC:L) as it requires only standard HTTP requests. Attack Requirements are None (AT:N). Privileges Required are Low (PR:L) as any authenticated non-admin user can exploit this. User Interaction is None (UI:N). Vulnerable Component Impact is None (VC:N) as the server itself is not compromised. Vulnerable Component Integrity is High (VI:H) because sensitive secrets and logs are exposed, which can lead to further compromise. Vulnerable Component Availability is None (VA:N). Subsequent Component Impact is None (SC:N, SI:N, SA:N) as the direct impact is limited to information disclosure within the vulnerable component, though downstream impacts are possible but not directly measured by the base vector.",
                "fixSummary": "The vulnerability is remediated by enforcing role-based access control on data serialization and template rendering. Non-admin users are now restricted to receiving only public enable flags and do not have access to webhook URLs, secrets, or delivery logs. The shared alert serializer explicitly removes sensitive fields for non-admin users, ensuring that both the notification page and the alerts API preserve administrative diagnostics while protecting sensitive data from unauthorized access.",
                "generatedAt": "2026-11-20T10:47:41.692593Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 3,
                      "confidence": "high",
                      "model": "qwen3.8:27b",
                      "score": 10
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "ee51a7b891c2ec45c5c6526b8a6ac91b2026442d6662c5e5888f49a807834861",
                "patchSummary": "The patch modifies app/alerts/alerts_core.py and app/notification/notification.py. In alerts_core.py, the alerts_to_json function now checks if the user is an admin and removes the webhook_url field from the JSON response for non-admin users. In notification.py, the index route now checks the current user\u0027s admin status. For non-admins, it passes a restricted configuration object containing only public enable flags (CASE_CREATE_ALERT_ENABLED, WEBHOOK_ENABLED) instead of the full ConfigModule, and it passes an empty list for alert_logs instead of the actual log data. This ensures sensitive configuration and logs are not embedded in the template context for unauthorized users.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "1445157891add7c04a63e0014cd96bf9802cac66",
                    "date": "Fri, 30 Oct 2026 14:18:20 +0100",
                    "patchSha256": "ee51a7b891c2ec45c5c6526b8a6ac91b2026442d6662c5e5888f49a807834861",
                    "retrievedAt": "2026-11-20T10:47:11.460773Z",
                    "source": "https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66.patch",
                    "subject": "fix: [security] Alert secrets and webhook logs stay"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T10:47:41Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Alert secrets and webhook logs stay",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-200",
                    "rationale": "The application exposed sensitive configuration data (webhook URLs, secrets, logs) to authenticated users who did not have the necessary privileges to view them."
                  },
                  {
                    "cweId": "CWE-862",
                    "rationale": "The application failed to enforce authorization checks when serializing alert data and rendering the notification page, allowing non-admin users to access admin-only data."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (1445157891add7c04a63e0014cd96bf9802cac66): fix: [security] Alert secrets and webhook logs stay",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66.patch"
                    ],
                    "timestamp": "2026-10-30T13:18:20Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66.patch"
                    ],
                    "timestamp": "2026-11-20T10:47:11.460773Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20405"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133665",
    "datePublished": "2026-11-20T10:58:14.508Z",
    "dateReserved": "2026-11-20T10:58:12.873Z",
    "dateUpdated": "2026-11-20T18:26:03.887Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133663 (GCVE-0-2026-133663)

Vulnerability from cvelistv5 – Published: 2026-11-20 10:58 – Updated: 2026-11-20 18:25
VLAI
Title
flowintel Insecure Direct Object Reference Allows Cross-Case Data Access and Modification
Summary
flowintel contains an Insecure Direct Object Reference (IDOR) vulnerability affecting multiple endpoints that manage child records of cases and tasks. An authenticated user with access to one case or task could read or modify records belonging to other cases by supplying identifiers of those records directly in the request. Affected operations include: - Reading task notes, subtasks, and URL tools by referencing their IDs without verifying they belong to the authorized task - Retrieving task assignees for a task belonging to a different case - Editing or deleting timeline events and timeline event links belonging to another case - Editing or removing case connectors belonging to another case - Invoking connector module operations (MISP, generic connectors, worker calls) by combining an authorized case with a connector link from a different case The root cause is that the application authorized the parent resource (case or task) in the URL but then fetched or operated on the child record independently without verifying the child's ownership relationship to the authorized parent. Security impact: an authenticated user can exfiltrate sensitive threat intelligence data from cases they do not have access to, and can modify or delete timeline entries, connector configurations, and other records in unauthorized cases, leading to data integrity loss and potential disruption of incident response workflows. Affected versions: flowintel prior to 4.1.0
SSVC
Exploitation: none Automatable: yes Technical Impact: total
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 10:47 UTC
Exploitation: none Automatable: no Technical Impact: total
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:24 UTC
CWE
  • CWE-639 - Authorization Bypass Through User-Controlled Key
  • CWE-862 - Missing Authorization
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 10:47
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae.patch 5d78368cb0dd…
Confidence
high
Commit Subject Patch SHA-256
fecba79506de fix: [security] Child records stay bound to their authorised 5d78368cb0dd…
Fix summary

The fix enforces ownership validation by requiring the authorized parent case or task ID in every core method that reads or writes child records. Before returning data or performing writes, each child record's case_id or task_id is explicitly compared against the resolved parent. Mismatched records are rejected with a 404 or 403 response. Connector module context building now verifies connector-to-case ownership before preparing module data, resolving credentials, or executing sync operations. Instance visibility checks are applied consistently across both web and API routes.

Patch summary

Added case_id or task_id ownership checks in CaseCore.py methods (remove_connector, edit_connector, edit_timeline_event, edit_timeline_event_link, _build_case_connector_module_context) by passing the authorized parent ID and comparing it against the child record's foreign key. Updated case.py get_assigned_users to verify task.case_id matches the resolved case. Updated case_api.py EditConnectorsCase and RemoveConnectors to validate connector ownership and instance visibility before delegating. Updated case_misp.py timeline and connector endpoints to pass case.id to core methods. Updated task_api.py note, subtask, and URL-tool GET handlers to verify the child's task_id matches the authorized task.

CVSS rationale

AV:N - exploitable over the network via HTTP API. AC:L - the attacker only needs to know or enumerate valid child record IDs; no special race conditions or complex setup required. AT:N - no attack target manipulation needed. PR:L - requires an authenticated user with access to at least one case or task. UI:N - no user interaction required. VC:H - full read access to sensitive threat intelligence data in other cases. VI:H - ability to modify or delete timeline events, links, and connectors in other cases. VA:N - no denial-of-service impact. SC/SI/SA:N - impact is confined to the flowintel application's own data store; no external system impact is evidenced.

Weakness rationale
  • CWE-639 The vulnerability is a classic IDOR: user-supplied identifiers for child records (notes, subtasks, connectors, timeline events) are used to fetch or modify objects without verifying the object belongs to the authorized parent case or task. The authorization check on the parent does not constrain the child lookup.
  • CWE-862 Several endpoints performed the operation on a child record without any ownership or authorization check tying the record to the user's authorized case, effectively missing the authorization step for the specific object being accessed.
Attack pattern rationale
  • CAPEC-1 The attacker manipulates URL path parameters or query string identifiers (note_id, subtask_id, url_tool_id, connector instance ID, timeline event ID) to reference objects belonging to other cases or tasks. The application did not validate that the referenced object belonged to the authorized parent, allowing cross-tenant data access. This is the closest CAPEC pattern; the attack does not involve protocol-level tampering but rather parameter substitution in REST-style URLs, which falls under the broader 'tamper with URLs' category.
Assumptions to verify
  • The affected version range is inferred from the explicit context (affected_version: 0, fixed_version: 4.1.0); no earlier version boundary is confirmed.
  • PR:L assumes the attacker needs at least one valid case or task membership; the patch does not clarify whether anonymous access is possible.
  • CAPEC-1 (Tamper with URLs) is the closest available pattern; the actual attack is parameter substitution in REST endpoints rather than classic URL tampering, but no more specific CAPEC exists for IDOR via REST identifiers.
  • VC:H and VI:H assume the data in other cases is sensitive and that modification/deletion constitutes full integrity loss; the patch does not specify data classification.
  • The patch does not indicate whether the vulnerability was disclosed publicly or under embargo; exploitation status is assumed to be 'none' based on absence of evidence.
  • Instance visibility enforcement (can_interact_with_case_connector) was added to the API route in case_api.py but was already present in case_misp.py; the patch ensures consistency rather than introducing a new check in all locations.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 7 11 high 6
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-10-30 13:58 UTC Fix developed Corrective change authored (fecba79506dea2886118fe5cd686d4edfdef60ae): fix: [security] Child records stay bound to their authorised https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae.patch
  2. 2026-11-20 10:46 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133663",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "total"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:24:45.904432Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:25:02.848Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/case/CaseCore.py",
            "app/case/case.py",
            "app/case/case_api.py",
            "app/case/case_misp.py",
            "app/case/task_api.py"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "flowintel contains an Insecure Direct Object Reference (IDOR) vulnerability affecting multiple endpoints that manage child records of cases and tasks. An authenticated user with access to one case or task could read or modify records belonging to other cases by supplying identifiers of those records directly in the request.\n\nAffected operations include:\n- Reading task notes, subtasks, and URL tools by referencing their IDs without verifying they belong to the authorized task\n- Retrieving task assignees for a task belonging to a different case\n- Editing or deleting timeline events and timeline event links belonging to another case\n- Editing or removing case connectors belonging to another case\n- Invoking connector module operations (MISP, generic connectors, worker calls) by combining an authorized case with a connector link from a different case\n\nThe root cause is that the application authorized the parent resource (case or task) in the URL but then fetched or operated on the child record independently without verifying the child\u0027s ownership relationship to the authorized parent.\n\nSecurity impact: an authenticated user can exfiltrate sensitive threat intelligence data from cases they do not have access to, and can modify or delete timeline entries, connector configurations, and other records in unauthorized cases, leading to data integrity loss and potential disruption of incident response workflows.\n\nAffected versions: flowintel prior to 4.1.0"
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-1",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-1 Tamper with URLs"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 8.6,
            "baseSeverity": "HIGH",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:N/SC:N/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "HIGH",
            "vulnIntegrityImpact": "HIGH",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "total"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T10:47:07Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-639",
              "description": "CWE-639 Authorization Bypass Through User-Controlled Key",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-862",
              "description": "CWE-862 Missing Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T10:58:03.710Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The fix enforces ownership validation by requiring the authorized parent case or task ID in every core method that reads or writes child records. Before returning data or performing writes, each child record\u0027s case_id or task_id is explicitly compared against the resolved parent. Mismatched records are rejected with a 404 or 403 response. Connector module context building now verifies connector-to-case ownership before preparing module data, resolving credentials, or executing sync operations. Instance visibility checks are applied consistently across both web and API routes."
        }
      ],
      "title": "flowintel Insecure Direct Object Reference Allows Cross-Case Data Access and Modification",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected version range is inferred from the explicit context (affected_version: 0, fixed_version: 4.1.0); no earlier version boundary is confirmed.",
                  "PR:L assumes the attacker needs at least one valid case or task membership; the patch does not clarify whether anonymous access is possible.",
                  "CAPEC-1 (Tamper with URLs) is the closest available pattern; the actual attack is parameter substitution in REST endpoints rather than classic URL tampering, but no more specific CAPEC exists for IDOR via REST identifiers.",
                  "VC:H and VI:H assume the data in other cases is sensitive and that modification/deletion constitutes full integrity loss; the patch does not specify data classification.",
                  "The patch does not indicate whether the vulnerability was disclosed publicly or under embargo; exploitation status is assumed to be \u0027none\u0027 based on absence of evidence.",
                  "Instance visibility enforcement (can_interact_with_case_connector) was added to the API route in case_api.py but was already present in case_misp.py; the patch ensures consistency rather than introducing a new check in all locations."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-1",
                    "rationale": "The attacker manipulates URL path parameters or query string identifiers (note_id, subtask_id, url_tool_id, connector instance ID, timeline event ID) to reference objects belonging to other cases or tasks. The application did not validate that the referenced object belonged to the authorized parent, allowing cross-tenant data access. This is the closest CAPEC pattern; the attack does not involve protocol-level tampering but rather parameter substitution in REST-style URLs, which falls under the broader \u0027tamper with URLs\u0027 category."
                  }
                ],
                "commit": "fecba79506dea2886118fe5cd686d4edfdef60ae",
                "confidence": "high",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N - exploitable over the network via HTTP API. AC:L - the attacker only needs to know or enumerate valid child record IDs; no special race conditions or complex setup required. AT:N - no attack target manipulation needed. PR:L - requires an authenticated user with access to at least one case or task. UI:N - no user interaction required. VC:H - full read access to sensitive threat intelligence data in other cases. VI:H - ability to modify or delete timeline events, links, and connectors in other cases. VA:N - no denial-of-service impact. SC/SI/SA:N - impact is confined to the flowintel application\u0027s own data store; no external system impact is evidenced.",
                "fixSummary": "The fix enforces ownership validation by requiring the authorized parent case or task ID in every core method that reads or writes child records. Before returning data or performing writes, each child record\u0027s case_id or task_id is explicitly compared against the resolved parent. Mismatched records are rejected with a 404 or 403 response. Connector module context building now verifies connector-to-case ownership before preparing module data, resolving credentials, or executing sync operations. Instance visibility checks are applied consistently across both web and API routes.",
                "generatedAt": "2026-11-20T10:47:07.135346Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 6,
                      "confidence": "high",
                      "model": "qwen3.8:27b",
                      "score": 7
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "5d78368cb0ddd8fd9dab90b7c709b83b5aff34787ec2c25e7ba984947f4454c9",
                "patchSummary": "Added case_id or task_id ownership checks in CaseCore.py methods (remove_connector, edit_connector, edit_timeline_event, edit_timeline_event_link, _build_case_connector_module_context) by passing the authorized parent ID and comparing it against the child record\u0027s foreign key. Updated case.py get_assigned_users to verify task.case_id matches the resolved case. Updated case_api.py EditConnectorsCase and RemoveConnectors to validate connector ownership and instance visibility before delegating. Updated case_misp.py timeline and connector endpoints to pass case.id to core methods. Updated task_api.py note, subtask, and URL-tool GET handlers to verify the child\u0027s task_id matches the authorized task.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "fecba79506dea2886118fe5cd686d4edfdef60ae",
                    "date": "Fri, 30 Oct 2026 14:58:13 +0100",
                    "patchSha256": "5d78368cb0ddd8fd9dab90b7c709b83b5aff34787ec2c25e7ba984947f4454c9",
                    "retrievedAt": "2026-11-20T10:46:36.448315Z",
                    "source": "https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae.patch",
                    "subject": "fix: [security] Child records stay bound to their authorised"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "total"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T10:47:07Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Child records stay bound to their authorised",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-639",
                    "rationale": "The vulnerability is a classic IDOR: user-supplied identifiers for child records (notes, subtasks, connectors, timeline events) are used to fetch or modify objects without verifying the object belongs to the authorized parent case or task. The authorization check on the parent does not constrain the child lookup."
                  },
                  {
                    "cweId": "CWE-862",
                    "rationale": "Several endpoints performed the operation on a child record without any ownership or authorization check tying the record to the user\u0027s authorized case, effectively missing the authorization step for the specific object being accessed."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (fecba79506dea2886118fe5cd686d4edfdef60ae): fix: [security] Child records stay bound to their authorised",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae.patch"
                    ],
                    "timestamp": "2026-10-30T13:58:13Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae.patch"
                    ],
                    "timestamp": "2026-11-20T10:46:36.448315Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20404"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133663",
    "datePublished": "2026-11-20T10:58:03.710Z",
    "dateReserved": "2026-11-20T10:58:01.964Z",
    "dateUpdated": "2026-11-20T18:25:02.848Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133651 (GCVE-0-2026-133651)

Vulnerability from cvelistv5 – Published: 2026-11-20 10:24 – Updated: 2026-11-20 18:23
VLAI
Title
FlowIntel CSRF via GET-accepted destructive web routes bypassing Flask-WTF protection
Summary
FlowIntel contained a Cross-Site Request Forgery (CSRF) vulnerability affecting fourteen web routes that performed destructive or state-changing operations. The affected routes accepted HTTP GET requests in addition to (or instead of) POST, which bypassed the Flask-WTF CSRF token validation that is only enforced on non-GET requests. The vulnerable operations included: - Case deletion and case merging - Organisation removal from a case - Task deletion (case and template) - Case template and note template deletion - Connector, connector instance, and icon deletion - Custom tag deletion and status change - Taxonomy and galaxy status changes An attacker who could induce an authenticated user to follow a crafted link or load a page containing a cross-site navigation (e.g., an image tag or redirect) could trigger any of these destructive actions using the victim's session cookie, without the victim's explicit consent or awareness. No special privileges beyond a valid authenticated session with the appropriate role (editor or admin) were required on the victim's account. Security impact: unauthorized deletion or modification of cases, tasks, templates, connectors, icons, custom tags, taxonomies, and galaxies belonging to the victim or the organisation, resulting in loss of data integrity. Affected versions: FlowIntel prior to 4.1.0
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 10:16 UTC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:23 UTC
CWE
  • CWE-352 - Cross-Site Request Forgery (CSRF)
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 10:16
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc.patch 8d58806814d8…
Confidence
high
Commit Subject Patch SHA-256
c29a27c4973f fix: [security] Require CSRF-protected POST for destructive 8d58806814d8…
Fix summary

The vulnerability is remediated by restricting all fourteen affected web routes to accept only POST requests, ensuring that Flask-WTF's CSRF token validation is enforced. All fifteen client-side callers (JavaScript fetch calls and HTML template scripts) were updated to send POST requests with the page's CSRF token in the X-CSRFToken header, preserving legitimate user interactions while blocking cross-site forged requests.

Patch summary

Backend: Changed the methods parameter on fourteen Flask route decorators from methods=['GET'] or methods=['GET','POST'] to methods=['POST'] across app/admin/admin.py, app/case/case.py, app/case/task.py, app/connectors/connectors.py, app/custom_tags/custom_tags.py, app/templating/templating.py, and app/tools/tools.py. Frontend: Updated fifteen fetch() calls in JavaScript files and HTML templates to include method: 'POST' and headers: {'X-CSRFToken': $('#csrf_token').val()}. No changes to permission decorators, query parameters, or API-key-authenticated routes.

CVSS rationale

AV:N - exploit is over the network via a web browser. AC:L - crafting a link or page that triggers a GET request is trivial. AT:N - no attack target manipulation beyond the victim's browser. PR:N - the attacker requires no account; the victim's existing session is used. UI:A - the victim must actively navigate to or click a crafted link (commit message references 'cross-site navigation'). VC:N, VI:N, VA:N - no direct impact on the application's own confidentiality, integrity, or availability. SC:N, SI:H, SA:N - the impact is on the integrity of the victim's data (deletion/modification of cases, tasks, templates, connectors, tags, taxonomies, galaxies) with no cross-system impact.

Weakness rationale
  • CWE-352 The routes accepted GET requests, which are not subject to Flask-WTF's CSRF token check. This allowed a cross-site navigation to invoke destructive state-changing operations using the victim's authenticated session, a textbook CSRF scenario.
Attack pattern rationale
  • CAPEC-124 The attack pattern involves an attacker crafting a request (in this case a GET navigation) that is sent to the target application using the victim's authenticated session cookie, causing the application to perform an unintended state-changing action. This is the closest and most direct CAPEC match for a GET-based CSRF bypass of token validation.
Assumptions to verify
  • The affected_version '0' from explicit context is treated as a placeholder indicating all versions prior to 4.1.0; the exact earliest affected version is not specified in the patch.
  • UI:A is selected based on the commit message referencing 'cross-site navigation'; if an attacker could use a passive vector (e.g., an auto-loading image tag), UI could be P, but the evidence specifically describes navigation.
  • PR:N assumes the attacker does not need their own account; the victim's authenticated session is sufficient. If the application requires the victim to have editor or admin role, this is a precondition on the victim, not a privilege requirement for the attacker.
  • CAPEC-124 is the closest standard mapping; no more specific CAPEC exists for GET-based CSRF bypass of framework token validation.
  • The patch does not indicate whether the CSRF token was previously absent from the page or only not enforced on GET; the fix implies the token was available in the DOM but the GET method bypassed validation.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 8 11 high 5
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-02 14:03 UTC Fix developed Corrective change authored (c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc): fix: [security] Require CSRF-protected POST for destructive https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc.patch
  2. 2026-11-20 10:15 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133651",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:23:00.226813Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:23:19.024Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/admin/admin.py",
            "app/case/case.py",
            "app/case/task.py",
            "app/connectors/connectors.py",
            "app/custom_tags/custom_tags.py",
            "app/templating/templating.py",
            "app/tools/tools.py"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "FlowIntel contained a Cross-Site Request Forgery (CSRF) vulnerability affecting fourteen web routes that performed destructive or state-changing operations. The affected routes accepted HTTP GET requests in addition to (or instead of) POST, which bypassed the Flask-WTF CSRF token validation that is only enforced on non-GET requests.\n\nThe vulnerable operations included:\n- Case deletion and case merging\n- Organisation removal from a case\n- Task deletion (case and template)\n- Case template and note template deletion\n- Connector, connector instance, and icon deletion\n- Custom tag deletion and status change\n- Taxonomy and galaxy status changes\n\nAn attacker who could induce an authenticated user to follow a crafted link or load a page containing a cross-site navigation (e.g., an image tag or redirect) could trigger any of these destructive actions using the victim\u0027s session cookie, without the victim\u0027s explicit consent or awareness. No special privileges beyond a valid authenticated session with the appropriate role (editor or admin) were required on the victim\u0027s account.\n\nSecurity impact: unauthorized deletion or modification of cases, tasks, templates, connectors, icons, custom tags, taxonomies, and galaxies belonging to the victim or the organisation, resulting in loss of data integrity.\n\nAffected versions: FlowIntel prior to 4.1.0"
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-124",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-124 Cross-Site Request Forgery"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 6.1,
            "baseSeverity": "MEDIUM",
            "privilegesRequired": "NONE",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "HIGH",
            "userInteraction": "ACTIVE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:A/VC:N/VI:N/VA:N/SC:N/SI:H/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "NONE",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T10:16:24Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-352",
              "description": "CWE-352 Cross-Site Request Forgery (CSRF)",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T10:24:17.885Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The vulnerability is remediated by restricting all fourteen affected web routes to accept only POST requests, ensuring that Flask-WTF\u0027s CSRF token validation is enforced. All fifteen client-side callers (JavaScript fetch calls and HTML template scripts) were updated to send POST requests with the page\u0027s CSRF token in the X-CSRFToken header, preserving legitimate user interactions while blocking cross-site forged requests."
        }
      ],
      "title": "FlowIntel CSRF via GET-accepted destructive web routes bypassing Flask-WTF protection",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected_version \u00270\u0027 from explicit context is treated as a placeholder indicating all versions prior to 4.1.0; the exact earliest affected version is not specified in the patch.",
                  "UI:A is selected based on the commit message referencing \u0027cross-site navigation\u0027; if an attacker could use a passive vector (e.g., an auto-loading image tag), UI could be P, but the evidence specifically describes navigation.",
                  "PR:N assumes the attacker does not need their own account; the victim\u0027s authenticated session is sufficient. If the application requires the victim to have editor or admin role, this is a precondition on the victim, not a privilege requirement for the attacker.",
                  "CAPEC-124 is the closest standard mapping; no more specific CAPEC exists for GET-based CSRF bypass of framework token validation.",
                  "The patch does not indicate whether the CSRF token was previously absent from the page or only not enforced on GET; the fix implies the token was available in the DOM but the GET method bypassed validation."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-124",
                    "rationale": "The attack pattern involves an attacker crafting a request (in this case a GET navigation) that is sent to the target application using the victim\u0027s authenticated session cookie, causing the application to perform an unintended state-changing action. This is the closest and most direct CAPEC match for a GET-based CSRF bypass of token validation."
                  }
                ],
                "commit": "c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc",
                "confidence": "high",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N - exploit is over the network via a web browser. AC:L - crafting a link or page that triggers a GET request is trivial. AT:N - no attack target manipulation beyond the victim\u0027s browser. PR:N - the attacker requires no account; the victim\u0027s existing session is used. UI:A - the victim must actively navigate to or click a crafted link (commit message references \u0027cross-site navigation\u0027). VC:N, VI:N, VA:N - no direct impact on the application\u0027s own confidentiality, integrity, or availability. SC:N, SI:H, SA:N - the impact is on the integrity of the victim\u0027s data (deletion/modification of cases, tasks, templates, connectors, tags, taxonomies, galaxies) with no cross-system impact.",
                "fixSummary": "The vulnerability is remediated by restricting all fourteen affected web routes to accept only POST requests, ensuring that Flask-WTF\u0027s CSRF token validation is enforced. All fifteen client-side callers (JavaScript fetch calls and HTML template scripts) were updated to send POST requests with the page\u0027s CSRF token in the X-CSRFToken header, preserving legitimate user interactions while blocking cross-site forged requests.",
                "generatedAt": "2026-11-20T10:16:24.400257Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 5,
                      "confidence": "high",
                      "model": "qwen3.8:27b",
                      "score": 8
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "8d58806814d8ac1f104c7206ff3b4c2e842c8a0c94205dcf816813439ac98863",
                "patchSummary": "Backend: Changed the methods parameter on fourteen Flask route decorators from methods=[\u0027GET\u0027] or methods=[\u0027GET\u0027,\u0027POST\u0027] to methods=[\u0027POST\u0027] across app/admin/admin.py, app/case/case.py, app/case/task.py, app/connectors/connectors.py, app/custom_tags/custom_tags.py, app/templating/templating.py, and app/tools/tools.py. Frontend: Updated fifteen fetch() calls in JavaScript files and HTML templates to include method: \u0027POST\u0027 and headers: {\u0027X-CSRFToken\u0027: $(\u0027#csrf_token\u0027).val()}. No changes to permission decorators, query parameters, or API-key-authenticated routes.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc",
                    "date": "Mon, 2 Nov 2026 15:03:14 +0100",
                    "patchSha256": "8d58806814d8ac1f104c7206ff3b4c2e842c8a0c94205dcf816813439ac98863",
                    "retrievedAt": "2026-11-20T10:15:57.346441Z",
                    "source": "https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc.patch",
                    "subject": "fix: [security] Require CSRF-protected POST for destructive"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T10:16:24Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Require CSRF-protected POST for destructive",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-352",
                    "rationale": "The routes accepted GET requests, which are not subject to Flask-WTF\u0027s CSRF token check. This allowed a cross-site navigation to invoke destructive state-changing operations using the victim\u0027s authenticated session, a textbook CSRF scenario."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc): fix: [security] Require CSRF-protected POST for destructive",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc.patch"
                    ],
                    "timestamp": "2026-11-02T14:03:14Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc.patch"
                    ],
                    "timestamp": "2026-11-20T10:15:57.346441Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20237"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133651",
    "datePublished": "2026-11-20T10:24:17.885Z",
    "dateReserved": "2026-11-20T10:24:15.465Z",
    "dateUpdated": "2026-11-20T18:23:19.024Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133648 (GCVE-0-2026-133648)

Vulnerability from cvelistv5 – Published: 2026-11-20 10:23 – Updated: 2026-11-20 18:22
VLAI
Title
Flowintel Disables TLS Certificate Verification for MISP and Mail Connections
Summary
Flowintel disabled TLS certificate verification for all MISP API connections and used unverified SSL contexts for IMAP and SMTP STARTTLS mail connections. This allowed an on-path (man-in-the-middle) attacker positioned between Flowintel and its MISP or mail servers to impersonate the remote server and intercept sensitive data. Affected connections include: - All MISP HTTPS API calls (event retrieval, attribute search, event creation, connectivity checks, and API validation) - SMTP STARTTLS connections used for sending email notifications - IMAP TLS connections used for fetching mail archives and testing connectivity An attacker who can observe or intercept network traffic between Flowintel and the MISP or mail server could: - Capture MISP API keys and mail credentials (SMTP/IMAP passwords) - Read all threat intelligence data exchanged (events, attributes, objects, reports) - Inject or modify threat intelligence data returned to Flowintel The vulnerability required no authentication on the part of the attacker and no user interaction; it was exploitable by any party with a network position between Flowintel and its configured MISP or mail endpoints. Affected versions: Flowintel prior to 4.1.0.
SSVC
Exploitation: none Automatable: yes Technical Impact: total
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 10:15 UTC
Exploitation: none Automatable: yes Technical Impact: total
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:22 UTC
CWE
  • CWE-295 - Improper Certificate Validation
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 10:15
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f.patch 286e145e7697…
Confidence
high
Commit Subject Patch SHA-256
b8e257ab1763 fix: [security] Verify TLS certificates for MISP and mail 286e145e7697…
Fix summary

The fix introduces a shared TLS utility module that enforces certificate and hostname verification by default for all MISP and mail connections. MISP clients now receive a proper SSL verification context instead of ssl=False, and IMAP/SMTP connections use a verified ssl.SSLContext. Operators may configure a private CA bundle path or explicitly disable verification (which fails closed on malformed input). The previous urllib3 warning suppression is removed, and upgrade documentation warns that previously accepted self-signed or mismatched certificates will now cause connection failures.

Patch summary

Added new file app/utils/tls.py providing misp_ssl() and mail_ssl_context() helpers that read MISP_VERIFY_SSL and MAIL_VERIFY_SSL config values (default True) and return appropriate verification settings or SSL contexts. Replaced all ssl=False arguments in PyMISP constructor calls across 8 files with ssl=misp_ssl(). Replaced bare server.starttls() with server.starttls(context=mail_ssl_context()) and added ssl_context=mail_ssl_context() to IMAP4_SSL calls. Removed all urllib3.disable_warnings() calls. Added MISP_VERIFY_SSL and MAIL_VERIFY_SSL to the default config template. Added installation documentation for TLS configuration and upgrade notes. The _verify_setting() helper raises ValueError on invalid input, ensuring fail-closed behavior.

CVSS rationale

AV:N - The attack is performed over the network by intercepting traffic between Flowintel and MISP/mail servers. AC:L - No special race conditions or complex setup are needed; the attacker only needs a network position on the path. AT:N - Flowintel is a server application; no active user on the target is required for the connection to be made. PR:N - The attacker does not need any credentials or account on Flowintel. UI:N - No user interaction is required. VC:H - All data in transit (API keys, mail credentials, full threat intelligence content) is readable by the attacker. VI:H - The attacker can modify or inject threat intelligence data returned to Flowintel. VA:N - No denial-of-service impact is indicated. SC:H - The MISP instance's data and the mail server's data are exposed to the attacker. SI:H - Forged or modified data can be injected into the MISP instance via the compromised channel. SA:N - No security authority impact is evident.

Weakness rationale
  • CWE-295 The application explicitly disabled TLS certificate verification by passing ssl=False to all MISP client connections and by using unverified SSL contexts for IMAP/SMTP. This is a textbook case of improper certificate validation where the client accepts any server certificate without verifying the chain or hostname.
Attack pattern rationale
  • CAPEC-119 The disabled certificate verification directly enables a classic man-in-the-middle attack: an on-path attacker presents a self-signed or rogue certificate, Flowintel accepts it without validation, and the attacker reads and modifies all traffic. CAPEC-119 is the closest standard attack pattern describing interception and modification of communications between two parties. The patch commit message explicitly references an 'on-path attacker' impersonating the server, confirming this mapping.
Assumptions to verify
  • The affected version range is inferred from the explicit context (affected_version '0', fixed_version '4.1.0'); the patch itself does not enumerate all prior releases.
  • CVSS VI:H assumes the attacker can successfully inject or modify threat intelligence data in transit; if the MISP API enforces additional integrity checks server-side, the practical integrity impact may be lower.
  • CAPEC-119 (Man-in-the-Middle) is selected as the closest standard attack pattern; the patch does not reference a specific CAPEC, and the attack could also be described under CAPEC-123 (Session Hijacking) or CAPEC-107 (Web Service Cross-Site Request Forgery) in a broader sense, but CAPEC-119 most directly matches the described on-path interception scenario.
  • The patch assumes MISP connector URLs use HTTPS; if an operator configures a plain HTTP MISP URL, TLS verification is not applicable and the vulnerability does not manifest in the same way.
  • PR:N assumes the attacker does not need any Flowintel account; the vulnerability is in the outbound connection from Flowintel to MISP/mail, not in Flowintel's own authentication.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 8 11 high 5
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-02 14:48 UTC Fix developed Corrective change authored (b8e257ab17632eb37c912485cbd78ee7514b764f): fix: [security] Verify TLS certificates for MISP and mail https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f.patch
  2. 2026-11-20 10:15 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133648",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "yes"
                  },
                  {
                    "Technical Impact": "total"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:22:26.198046Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:22:41.449Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/connectors/connectors_core.py",
            "app/modules/notify_user/email.py",
            "app/modules/receive_from/receive_misp_object.py",
            "app/modules/send_to/misp_event.py",
            "app/modules/send_to/misp_event_task.py",
            "app/modules/send_to/misp_object_event.py",
            "app/tools/tools_core.py",
            "app/tools/validate_api.py"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "Flowintel disabled TLS certificate verification for all MISP API connections and used unverified SSL contexts for IMAP and SMTP STARTTLS mail connections. This allowed an on-path (man-in-the-middle) attacker positioned between Flowintel and its MISP or mail servers to impersonate the remote server and intercept sensitive data.\n\nAffected connections include:\n- All MISP HTTPS API calls (event retrieval, attribute search, event creation, connectivity checks, and API validation)\n- SMTP STARTTLS connections used for sending email notifications\n- IMAP TLS connections used for fetching mail archives and testing connectivity\n\nAn attacker who can observe or intercept network traffic between Flowintel and the MISP or mail server could:\n- Capture MISP API keys and mail credentials (SMTP/IMAP passwords)\n- Read all threat intelligence data exchanged (events, attributes, objects, reports)\n- Inject or modify threat intelligence data returned to Flowintel\n\nThe vulnerability required no authentication on the part of the attacker and no user interaction; it was exploitable by any party with a network position between Flowintel and its configured MISP or mail endpoints.\n\nAffected versions: Flowintel prior to 4.1.0."
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-119",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-119 Man-in-the-Middle"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 9.9,
            "baseSeverity": "CRITICAL",
            "privilegesRequired": "NONE",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "HIGH",
            "subIntegrityImpact": "HIGH",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:N/SC:H/SI:H/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "HIGH",
            "vulnIntegrityImpact": "HIGH",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "total"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T10:15:55Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-295",
              "description": "CWE-295 Improper Certificate Validation",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T10:23:53.537Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The fix introduces a shared TLS utility module that enforces certificate and hostname verification by default for all MISP and mail connections. MISP clients now receive a proper SSL verification context instead of ssl=False, and IMAP/SMTP connections use a verified ssl.SSLContext. Operators may configure a private CA bundle path or explicitly disable verification (which fails closed on malformed input). The previous urllib3 warning suppression is removed, and upgrade documentation warns that previously accepted self-signed or mismatched certificates will now cause connection failures."
        }
      ],
      "title": "Flowintel Disables TLS Certificate Verification for MISP and Mail Connections",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected version range is inferred from the explicit context (affected_version \u00270\u0027, fixed_version \u00274.1.0\u0027); the patch itself does not enumerate all prior releases.",
                  "CVSS VI:H assumes the attacker can successfully inject or modify threat intelligence data in transit; if the MISP API enforces additional integrity checks server-side, the practical integrity impact may be lower.",
                  "CAPEC-119 (Man-in-the-Middle) is selected as the closest standard attack pattern; the patch does not reference a specific CAPEC, and the attack could also be described under CAPEC-123 (Session Hijacking) or CAPEC-107 (Web Service Cross-Site Request Forgery) in a broader sense, but CAPEC-119 most directly matches the described on-path interception scenario.",
                  "The patch assumes MISP connector URLs use HTTPS; if an operator configures a plain HTTP MISP URL, TLS verification is not applicable and the vulnerability does not manifest in the same way.",
                  "PR:N assumes the attacker does not need any Flowintel account; the vulnerability is in the outbound connection from Flowintel to MISP/mail, not in Flowintel\u0027s own authentication."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-119",
                    "rationale": "The disabled certificate verification directly enables a classic man-in-the-middle attack: an on-path attacker presents a self-signed or rogue certificate, Flowintel accepts it without validation, and the attacker reads and modifies all traffic. CAPEC-119 is the closest standard attack pattern describing interception and modification of communications between two parties. The patch commit message explicitly references an \u0027on-path attacker\u0027 impersonating the server, confirming this mapping."
                  }
                ],
                "commit": "b8e257ab17632eb37c912485cbd78ee7514b764f",
                "confidence": "high",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N - The attack is performed over the network by intercepting traffic between Flowintel and MISP/mail servers. AC:L - No special race conditions or complex setup are needed; the attacker only needs a network position on the path. AT:N - Flowintel is a server application; no active user on the target is required for the connection to be made. PR:N - The attacker does not need any credentials or account on Flowintel. UI:N - No user interaction is required. VC:H - All data in transit (API keys, mail credentials, full threat intelligence content) is readable by the attacker. VI:H - The attacker can modify or inject threat intelligence data returned to Flowintel. VA:N - No denial-of-service impact is indicated. SC:H - The MISP instance\u0027s data and the mail server\u0027s data are exposed to the attacker. SI:H - Forged or modified data can be injected into the MISP instance via the compromised channel. SA:N - No security authority impact is evident.",
                "fixSummary": "The fix introduces a shared TLS utility module that enforces certificate and hostname verification by default for all MISP and mail connections. MISP clients now receive a proper SSL verification context instead of ssl=False, and IMAP/SMTP connections use a verified ssl.SSLContext. Operators may configure a private CA bundle path or explicitly disable verification (which fails closed on malformed input). The previous urllib3 warning suppression is removed, and upgrade documentation warns that previously accepted self-signed or mismatched certificates will now cause connection failures.",
                "generatedAt": "2026-11-20T10:15:55.134458Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 5,
                      "confidence": "high",
                      "model": "qwen3.8:27b",
                      "score": 8
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "286e145e7697fc6bd5fcba49ad4998a8b621fa9a18f2b287ba3c344f26f06338",
                "patchSummary": "Added new file app/utils/tls.py providing misp_ssl() and mail_ssl_context() helpers that read MISP_VERIFY_SSL and MAIL_VERIFY_SSL config values (default True) and return appropriate verification settings or SSL contexts. Replaced all ssl=False arguments in PyMISP constructor calls across 8 files with ssl=misp_ssl(). Replaced bare server.starttls() with server.starttls(context=mail_ssl_context()) and added ssl_context=mail_ssl_context() to IMAP4_SSL calls. Removed all urllib3.disable_warnings() calls. Added MISP_VERIFY_SSL and MAIL_VERIFY_SSL to the default config template. Added installation documentation for TLS configuration and upgrade notes. The _verify_setting() helper raises ValueError on invalid input, ensuring fail-closed behavior.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "b8e257ab17632eb37c912485cbd78ee7514b764f",
                    "date": "Mon, 2 Nov 2026 15:48:58 +0100",
                    "patchSha256": "286e145e7697fc6bd5fcba49ad4998a8b621fa9a18f2b287ba3c344f26f06338",
                    "retrievedAt": "2026-11-20T10:15:19.966655Z",
                    "source": "https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f.patch",
                    "subject": "fix: [security] Verify TLS certificates for MISP and mail"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "total"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T10:15:55Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Verify TLS certificates for MISP and mail",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-295",
                    "rationale": "The application explicitly disabled TLS certificate verification by passing ssl=False to all MISP client connections and by using unverified SSL contexts for IMAP/SMTP. This is a textbook case of improper certificate validation where the client accepts any server certificate without verifying the chain or hostname."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (b8e257ab17632eb37c912485cbd78ee7514b764f): fix: [security] Verify TLS certificates for MISP and mail",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f.patch"
                    ],
                    "timestamp": "2026-11-02T14:48:58Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f.patch"
                    ],
                    "timestamp": "2026-11-20T10:15:19.966655Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20254"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133648",
    "datePublished": "2026-11-20T10:23:53.537Z",
    "dateReserved": "2026-11-20T10:23:51.984Z",
    "dateUpdated": "2026-11-20T18:22:41.449Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133631 (GCVE-0-2026-133631)

Vulnerability from cvelistv5 – Published: 2026-11-20 09:10 – Updated: 2026-11-20 18:18
VLAI
Title
FlowIntel Insecure Note Variable Resolution, Over-Permissive Chatbot API Access, and Credential Exposure in Process Arguments
Summary
FlowIntel contains multiple security flaws in its note variable resolution, chatbot API tooling, and AI report generation. - Note variable references (e.g. cross-case or cross-task lookups) were resolved using the viewing user's privileges rather than the destination case's audience. An authenticated user with access to a shared or imported note could embed references to private cases, causing sensitive case data to be pulled into AI report requests, exports, and MISP sends without the referenced case's audience having granted read access. - The chatbot API tool permitted the language model to invoke any FlowIntel API endpoint, including administrative routes and state-changing GET operations, without an allow-list or path validation. An attacker who can influence the model's instructions (e.g. via untrusted case content) could induce data modification or access restricted endpoints. - AI report generation passed the Ollama bearer token and full case content as command-line arguments to a child process, exposing them in the system process table to any local user. Affected versions: FlowIntel prior to 4.1.0.
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 09:04 UTC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:17 UTC
CWE
  • CWE-862 - Missing Authorization
  • CWE-284 - Improper Access Control
  • CWE-200 - Exposure of Sensitive Information to an Unauthorized Actor
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 09:04
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1.patch c2ffe93f6a63…
Confidence
medium
Commit Subject Patch SHA-256
34d151b8dceb fix: [security] Keep note references and AI tooling within c2ffe93f6a63…
Fix summary

Three remediations are applied: (1) Note variable resolution now enforces an audience-based authorization check so that cross-case and cross-task references only expand when the destination case's assigned organisations can read the referenced case; (2) the chatbot API tool is restricted to an explicit allow-list of reviewed read-only and search operations, with administrative endpoints always blocked, path validation against injection and SSRF, and write operations gated behind an explicit CHATBOT_ALLOW_WRITES configuration flag; (3) AI report requests are delivered to the child process via a private anonymous temporary file on stdin instead of argv, keeping credentials and case content out of the process listing.

Patch summary

In note_variables.py, a new _audience_can_view() function verifies that the destination case's organisations are a subset of the referenced case's organisations before resolving cross-case or cross-task references; case_notes.py passes the resolved case object instead of a raw integer ID. In chatbot_core.py, a _chatbot_api_request() validator and a _READ_API_PATHS regex allow-list are added; unsupported methods, external URLs, path traversal, admin routes, and non-allow-listed operations are refused before the MCP call. In CaseCore.py, the subprocess invocation for AI reports is changed from passing URL, payload, and headers as argv to writing them as JSON into a tempfile.TemporaryFile and feeding it via stdin. A CHATBOT_ALLOW_WRITES flag is added to the default config, and documentation is updated in the installation and user manuals.

CVSS rationale

AV:N: FlowIntel is a network-accessible web application. AC:L: An authenticated user can craft note references or interact with the chatbot without special conditions. AT:N: No specific attack target is required. PR:L: A low-privilege authenticated user with access to a shared note or the chatbot is sufficient. UI:N: No victim interaction beyond normal use is needed. VC:H: Private case data (titles, notes, MISP objects, user data) is exposed to unauthorized viewers and external AI/MISP systems. VI:L: The chatbot can invoke state-changing endpoints, allowing limited data modification. VA:N: No availability impact is evidenced. SC:L: Sensitive data is forwarded to external AI (Ollama) and MISP instances. SI:N: No subsequent integrity impact is demonstrated. SA:N: No subsequent availability impact.

Weakness rationale
  • CWE-862 Note variable resolution did not verify that the destination case's audience had read access to the referenced case, allowing an authenticated user to pull private data from cases they are not authorised to view.
  • CWE-284 The chatbot API tool lacked an allow-list or restriction on which API endpoints the model could invoke, permitting access to administrative and state-changing routes.
  • CWE-200 The Ollama bearer token and case content were passed as command-line arguments, visible in the local process table to any user on the host.
Attack pattern rationale
  • CAPEC-126 The note variable resolution relied solely on the viewer's own privileges without checking the destination case's audience, effectively an incorrectly implemented authorization scheme. An attacker crafts cross-case references in a note they can view to exfiltrate data from cases outside their access scope. This is the closest CAPEC to the missing audience-based authorization check.
  • CAPEC-12 The chatbot API tool was configured to permit unrestricted API calls without an allow-list, representing an application configuration error that exposes administrative and write endpoints to model-driven requests. The fix introduces a restrictive default configuration. This mapping is approximate; no CAPEC specifically covers LLM-driven API over-permissiveness.
Assumptions to verify
  • The affected version range is inferred from the explicit context (affected_version 0, fixed_version 4.1.0); no specific prior version boundary is confirmed in the patch.
  • PR:L assumes the attacker is an authenticated FlowIntel user with access to at least one shared or imported note and the chatbot interface; unauthenticated access is not evidenced.
  • VI:L for the chatbot assumes CHATBOT_ALLOW_WRITES is disabled by default; if enabled, the integrity impact could be higher.
  • SC:L assumes AI report data and MISP sends reach external systems; if the deployment is fully air-gapped, SC could be N.
  • CAPEC-126 is the closest available pattern for the missing audience-based authorization; no CAPEC specifically models cross-entity reference resolution authorization gaps.
  • CAPEC-12 is used for the chatbot over-permissiveness; no CAPEC specifically covers LLM-driven unrestricted API invocation.
  • The credential-in-argv issue (CWE-200) is a local information exposure and may not be exploitable in containerized or single-user deployments; its contribution to the overall CVSS is minor.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 5 11 medium 7
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-03 10:53 UTC Fix developed Corrective change authored (34d151b8dceb1fea925b353f04a9483aba0a22b1): fix: [security] Keep note references and AI tooling within https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1.patch
  2. 2026-11-20 09:03 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133631",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:17:51.923396Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:18:08.558Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/case/CaseCore.py",
            "app/case/case_notes.py",
            "app/chatbot/chatbot_core.py",
            "app/utils/note_variables.py"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "FlowIntel contains multiple security flaws in its note variable resolution, chatbot API tooling, and AI report generation.\n\n- Note variable references (e.g. cross-case or cross-task lookups) were resolved using the viewing user\u0027s privileges rather than the destination case\u0027s audience. An authenticated user with access to a shared or imported note could embed references to private cases, causing sensitive case data to be pulled into AI report requests, exports, and MISP sends without the referenced case\u0027s audience having granted read access.\n\n- The chatbot API tool permitted the language model to invoke any FlowIntel API endpoint, including administrative routes and state-changing GET operations, without an allow-list or path validation. An attacker who can influence the model\u0027s instructions (e.g. via untrusted case content) could induce data modification or access restricted endpoints.\n\n- AI report generation passed the Ollama bearer token and full case content as command-line arguments to a child process, exposing them in the system process table to any local user.\n\nAffected versions: FlowIntel prior to 4.1.0."
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-126",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-126 Exploiting Incorrectly Implemented Authentication Schemes"
            }
          ]
        },
        {
          "capecId": "CAPEC-12",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-12 Application Configuration Error"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 7.1,
            "baseSeverity": "HIGH",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "LOW",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:L/VA:N/SC:L/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "HIGH",
            "vulnIntegrityImpact": "LOW",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T09:04:05Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-862",
              "description": "CWE-862 Missing Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-284",
              "description": "CWE-284 Improper Access Control",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-200",
              "description": "CWE-200 Exposure of Sensitive Information to an Unauthorized Actor",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T09:11:43.154Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "Three remediations are applied: (1) Note variable resolution now enforces an audience-based authorization check so that cross-case and cross-task references only expand when the destination case\u0027s assigned organisations can read the referenced case; (2) the chatbot API tool is restricted to an explicit allow-list of reviewed read-only and search operations, with administrative endpoints always blocked, path validation against injection and SSRF, and write operations gated behind an explicit CHATBOT_ALLOW_WRITES configuration flag; (3) AI report requests are delivered to the child process via a private anonymous temporary file on stdin instead of argv, keeping credentials and case content out of the process listing."
        }
      ],
      "title": "FlowIntel Insecure Note Variable Resolution, Over-Permissive Chatbot API Access, and Credential Exposure in Process Arguments",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected version range is inferred from the explicit context (affected_version 0, fixed_version 4.1.0); no specific prior version boundary is confirmed in the patch.",
                  "PR:L assumes the attacker is an authenticated FlowIntel user with access to at least one shared or imported note and the chatbot interface; unauthenticated access is not evidenced.",
                  "VI:L for the chatbot assumes CHATBOT_ALLOW_WRITES is disabled by default; if enabled, the integrity impact could be higher.",
                  "SC:L assumes AI report data and MISP sends reach external systems; if the deployment is fully air-gapped, SC could be N.",
                  "CAPEC-126 is the closest available pattern for the missing audience-based authorization; no CAPEC specifically models cross-entity reference resolution authorization gaps.",
                  "CAPEC-12 is used for the chatbot over-permissiveness; no CAPEC specifically covers LLM-driven unrestricted API invocation.",
                  "The credential-in-argv issue (CWE-200) is a local information exposure and may not be exploitable in containerized or single-user deployments; its contribution to the overall CVSS is minor."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-126",
                    "rationale": "The note variable resolution relied solely on the viewer\u0027s own privileges without checking the destination case\u0027s audience, effectively an incorrectly implemented authorization scheme. An attacker crafts cross-case references in a note they can view to exfiltrate data from cases outside their access scope. This is the closest CAPEC to the missing audience-based authorization check."
                  },
                  {
                    "capecId": "CAPEC-12",
                    "rationale": "The chatbot API tool was configured to permit unrestricted API calls without an allow-list, representing an application configuration error that exposes administrative and write endpoints to model-driven requests. The fix introduces a restrictive default configuration. This mapping is approximate; no CAPEC specifically covers LLM-driven API over-permissiveness."
                  }
                ],
                "commit": "34d151b8dceb1fea925b353f04a9483aba0a22b1",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N: FlowIntel is a network-accessible web application. AC:L: An authenticated user can craft note references or interact with the chatbot without special conditions. AT:N: No specific attack target is required. PR:L: A low-privilege authenticated user with access to a shared note or the chatbot is sufficient. UI:N: No victim interaction beyond normal use is needed. VC:H: Private case data (titles, notes, MISP objects, user data) is exposed to unauthorized viewers and external AI/MISP systems. VI:L: The chatbot can invoke state-changing endpoints, allowing limited data modification. VA:N: No availability impact is evidenced. SC:L: Sensitive data is forwarded to external AI (Ollama) and MISP instances. SI:N: No subsequent integrity impact is demonstrated. SA:N: No subsequent availability impact.",
                "fixSummary": "Three remediations are applied: (1) Note variable resolution now enforces an audience-based authorization check so that cross-case and cross-task references only expand when the destination case\u0027s assigned organisations can read the referenced case; (2) the chatbot API tool is restricted to an explicit allow-list of reviewed read-only and search operations, with administrative endpoints always blocked, path validation against injection and SSRF, and write operations gated behind an explicit CHATBOT_ALLOW_WRITES configuration flag; (3) AI report requests are delivered to the child process via a private anonymous temporary file on stdin instead of argv, keeping credentials and case content out of the process listing.",
                "generatedAt": "2026-11-20T09:04:05.830413Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 7,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 5
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "c2ffe93f6a63b984f6d1a733093060b3def8170a7f47b9662ca137faa3a5d7e4",
                "patchSummary": "In note_variables.py, a new _audience_can_view() function verifies that the destination case\u0027s organisations are a subset of the referenced case\u0027s organisations before resolving cross-case or cross-task references; case_notes.py passes the resolved case object instead of a raw integer ID. In chatbot_core.py, a _chatbot_api_request() validator and a _READ_API_PATHS regex allow-list are added; unsupported methods, external URLs, path traversal, admin routes, and non-allow-listed operations are refused before the MCP call. In CaseCore.py, the subprocess invocation for AI reports is changed from passing URL, payload, and headers as argv to writing them as JSON into a tempfile.TemporaryFile and feeding it via stdin. A CHATBOT_ALLOW_WRITES flag is added to the default config, and documentation is updated in the installation and user manuals.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "34d151b8dceb1fea925b353f04a9483aba0a22b1",
                    "date": "Tue, 3 Nov 2026 11:53:18 +0100",
                    "patchSha256": "c2ffe93f6a63b984f6d1a733093060b3def8170a7f47b9662ca137faa3a5d7e4",
                    "retrievedAt": "2026-11-20T09:03:27.928581Z",
                    "source": "https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1.patch",
                    "subject": "fix: [security] Keep note references and AI tooling within"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T09:04:05Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Keep note references and AI tooling within",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-862",
                    "rationale": "Note variable resolution did not verify that the destination case\u0027s audience had read access to the referenced case, allowing an authenticated user to pull private data from cases they are not authorised to view."
                  },
                  {
                    "cweId": "CWE-284",
                    "rationale": "The chatbot API tool lacked an allow-list or restriction on which API endpoints the model could invoke, permitting access to administrative and state-changing routes."
                  },
                  {
                    "cweId": "CWE-200",
                    "rationale": "The Ollama bearer token and case content were passed as command-line arguments, visible in the local process table to any user on the host."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (34d151b8dceb1fea925b353f04a9483aba0a22b1): fix: [security] Keep note references and AI tooling within",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1.patch"
                    ],
                    "timestamp": "2026-11-03T10:53:18Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1.patch"
                    ],
                    "timestamp": "2026-11-20T09:03:27.928581Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20355"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133631",
    "datePublished": "2026-11-20T09:10:23.474Z",
    "dateReserved": "2026-11-20T09:10:21.313Z",
    "dateUpdated": "2026-11-20T18:18:08.558Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133630 (GCVE-0-2026-133630)

Vulnerability from cvelistv5 – Published: 2026-11-20 09:09 – Updated: 2026-11-20 18:17
VLAI
Title
flowintel CI/CD Workflow Exposes Checkout Credentials in Docker Build Cache
Summary
The flowintel Docker build workflow used the GitHub Actions checkout action with default credential persistence, causing the job authentication token to be written into the repository's Git configuration file. This file became part of the Docker build context and could be retained in exported build cache layers, even though Git metadata was stripped from the final runtime image. An attacker with access to the CI/CD build cache or build environment could extract the persisted token and use it to gain unauthorized access to the private source repository. Impact: - Exposure of a valid authentication token to the private repository - Potential unauthorized read access to source code and repository metadata Affected versions: flowintel < 4.1.0
SSVC
Exploitation: none Automatable: no Technical Impact: partial
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 09:03 UTC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:16 UTC
CWE
  • CWE-522 - Insufficiently Protected Credentials
  • CWE-200 - Exposure of Sensitive Information to an Unauthorized Actor
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 09:03
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05.patch 6cdbee24e7fd…
Confidence
medium
Commit Subject Patch SHA-256
6b26b3f8cf9f fix: [security] Keep checkout credentials out of Docker build 6cdbee24e7fd…
Fix summary

The checkout step is configured to not persist credentials into the workspace, and the full branch history is fetched during the authenticated checkout phase. This ensures the job token is never written to the Git configuration file that enters the Docker build context, eliminating the credential from build cache layers while preserving the ability to resolve release tags against the main branch.

Patch summary

In .github/workflows/docker.yml, the actions/checkout@v4 step is updated with persist-credentials: false to prevent the job token from being stored in .git/config, and fetch-depth: 0 to retrieve full history during the authenticated checkout. The subsequent git fetch origin main command in the branch-check step is removed since the origin/main reference is already available locally from the full-depth checkout.

CVSS rationale

AV:L: Exploitation requires local access to the CI/CD build environment or exported build cache; it is not remotely exploitable over a network. AC:L: Once build cache access is obtained, extracting the token from .git/config is straightforward. AT:N: No active user interaction is required. PR:H: The attacker needs high privileges (access to the CI/CD build infrastructure or build cache storage). UI:N: No user interaction is needed. VC/VI/VA:N: The CI/CD system itself is not directly compromised. SC:H: The exposed token grants full read access to the private repository, a separate system. SI:N: The job token typically has read-only scope by default; no evidence of write access. SA:N: No availability impact on the secondary system.

Weakness rationale
  • CWE-522 The GitHub Actions job token (a bearer credential) was persisted in .git/config and included in the Docker build context and cache layers without being removed, constituting insufficient protection of credentials in a build artifact.
  • CWE-200 The credential was exposed in Docker build cache layers accessible to anyone with access to the build environment, constituting information disclosure of a sensitive authentication token.
Attack pattern rationale
  • CAPEC-1 The closest plausible CAPEC is Session Hijacking: the exposed job token functions as a bearer credential that, if obtained from the build cache, allows an attacker to impersonate the CI/CD system's authenticated session and access the private repository. This is an imperfect mapping because the token is a CI/CD job token rather than a user session cookie, but the underlying mechanism (stealing a valid credential to assume an authenticated identity) aligns with the CAPEC-1 pattern. No CAPEC specifically covers credential leakage in build artifacts, so CAPEC-1 is the best available match.
Assumptions to verify
  • The GitHub Actions job token has read-only scope to the repository by default; if write scope is configured, SI could be elevated to Low.
  • The build cache is accessible to a limited set of actors (CI/CD operators or platform administrators); broader exposure (e.g., public Docker registry with cached layers) would change AV and PR ratings.
  • CAPEC-1 (Session Hijacking) is the closest available mapping; no CAPEC specifically addresses credential leakage in CI/CD build artifacts.
  • The affected version range is inferred from the explicit context (affected_version: 0, fixed_version: 4.1.0); the exact first vulnerable release is not confirmed by the patch alone.
  • The vulnerability is specific to the GitHub Actions Docker build workflow and does not affect the flowintel application runtime.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 7 11 medium 5
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-03 12:17 UTC Fix developed Corrective change authored (6b26b3f8cf9fab13244d1496443f822114ce4d05): fix: [security] Keep checkout credentials out of Docker build https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05.patch
  2. 2026-11-20 09:02 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133630",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:16:56.116043Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:17:10.037Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            ".github/workflows/docker.yml"
          ],
          "product": "flowintel",
          "programFiles": [
            ".github/workflows/docker.yml"
          ],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "The flowintel Docker build workflow used the GitHub Actions checkout action with default credential persistence, causing the job authentication token to be written into the repository\u0027s Git configuration file. This file became part of the Docker build context and could be retained in exported build cache layers, even though Git metadata was stripped from the final runtime image.\n\nAn attacker with access to the CI/CD build cache or build environment could extract the persisted token and use it to gain unauthorized access to the private source repository.\n\nImpact:\n- Exposure of a valid authentication token to the private repository\n- Potential unauthorized read access to source code and repository metadata\n\nAffected versions: flowintel \u003c 4.1.0"
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-1",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-1 Session Hijacking"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "LOCAL",
            "baseScore": 5.9,
            "baseSeverity": "MEDIUM",
            "privilegesRequired": "HIGH",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "HIGH",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:L/AC:L/AT:N/PR:H/UI:N/VC:N/VI:N/VA:N/SC:H/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "NONE",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "no"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T09:03:23Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-522",
              "description": "CWE-522 Insufficiently Protected Credentials",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-200",
              "description": "CWE-200 Exposure of Sensitive Information to an Unauthorized Actor",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T09:09:40.890Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The checkout step is configured to not persist credentials into the workspace, and the full branch history is fetched during the authenticated checkout phase. This ensures the job token is never written to the Git configuration file that enters the Docker build context, eliminating the credential from build cache layers while preserving the ability to resolve release tags against the main branch."
        }
      ],
      "title": "flowintel CI/CD Workflow Exposes Checkout Credentials in Docker Build Cache",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The GitHub Actions job token has read-only scope to the repository by default; if write scope is configured, SI could be elevated to Low.",
                  "The build cache is accessible to a limited set of actors (CI/CD operators or platform administrators); broader exposure (e.g., public Docker registry with cached layers) would change AV and PR ratings.",
                  "CAPEC-1 (Session Hijacking) is the closest available mapping; no CAPEC specifically addresses credential leakage in CI/CD build artifacts.",
                  "The affected version range is inferred from the explicit context (affected_version: 0, fixed_version: 4.1.0); the exact first vulnerable release is not confirmed by the patch alone.",
                  "The vulnerability is specific to the GitHub Actions Docker build workflow and does not affect the flowintel application runtime."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-1",
                    "rationale": "The closest plausible CAPEC is Session Hijacking: the exposed job token functions as a bearer credential that, if obtained from the build cache, allows an attacker to impersonate the CI/CD system\u0027s authenticated session and access the private repository. This is an imperfect mapping because the token is a CI/CD job token rather than a user session cookie, but the underlying mechanism (stealing a valid credential to assume an authenticated identity) aligns with the CAPEC-1 pattern. No CAPEC specifically covers credential leakage in build artifacts, so CAPEC-1 is the best available match."
                  }
                ],
                "commit": "6b26b3f8cf9fab13244d1496443f822114ce4d05",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:L: Exploitation requires local access to the CI/CD build environment or exported build cache; it is not remotely exploitable over a network. AC:L: Once build cache access is obtained, extracting the token from .git/config is straightforward. AT:N: No active user interaction is required. PR:H: The attacker needs high privileges (access to the CI/CD build infrastructure or build cache storage). UI:N: No user interaction is needed. VC/VI/VA:N: The CI/CD system itself is not directly compromised. SC:H: The exposed token grants full read access to the private repository, a separate system. SI:N: The job token typically has read-only scope by default; no evidence of write access. SA:N: No availability impact on the secondary system.",
                "fixSummary": "The checkout step is configured to not persist credentials into the workspace, and the full branch history is fetched during the authenticated checkout phase. This ensures the job token is never written to the Git configuration file that enters the Docker build context, eliminating the credential from build cache layers while preserving the ability to resolve release tags against the main branch.",
                "generatedAt": "2026-11-20T09:03:23.235755Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 5,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 7
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "6cdbee24e7fdf8bec92a1ccaa19b9f8c7fbf5229080dfb4029456b9fe84cf493",
                "patchSummary": "In .github/workflows/docker.yml, the actions/checkout@v4 step is updated with persist-credentials: false to prevent the job token from being stored in .git/config, and fetch-depth: 0 to retrieve full history during the authenticated checkout. The subsequent git fetch origin main command in the branch-check step is removed since the origin/main reference is already available locally from the full-depth checkout.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "6b26b3f8cf9fab13244d1496443f822114ce4d05",
                    "date": "Tue, 3 Nov 2026 13:17:37 +0100",
                    "patchSha256": "6cdbee24e7fdf8bec92a1ccaa19b9f8c7fbf5229080dfb4029456b9fe84cf493",
                    "retrievedAt": "2026-11-20T09:02:44.223116Z",
                    "source": "https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05.patch",
                    "subject": "fix: [security] Keep checkout credentials out of Docker build"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "no"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T09:03:23Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Keep checkout credentials out of Docker build",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-522",
                    "rationale": "The GitHub Actions job token (a bearer credential) was persisted in .git/config and included in the Docker build context and cache layers without being removed, constituting insufficient protection of credentials in a build artifact."
                  },
                  {
                    "cweId": "CWE-200",
                    "rationale": "The credential was exposed in Docker build cache layers accessible to anyone with access to the build environment, constituting information disclosure of a sensitive authentication token."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (6b26b3f8cf9fab13244d1496443f822114ce4d05): fix: [security] Keep checkout credentials out of Docker build",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05.patch"
                    ],
                    "timestamp": "2026-11-03T12:17:37Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05.patch"
                    ],
                    "timestamp": "2026-11-20T09:02:44.223116Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20377"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133630",
    "datePublished": "2026-11-20T09:09:40.890Z",
    "dateReserved": "2026-11-20T09:09:38.574Z",
    "dateUpdated": "2026-11-20T18:17:10.037Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133627 (GCVE-0-2026-133627)

Vulnerability from cvelistv5 – Published: 2026-11-20 08:52 – Updated: 2026-11-20 18:16
VLAI
Title
flowintel Case History Log Injection via Line Separator Characters in User-Authored Fields
Summary
flowintel contains a log injection vulnerability in its case history (audit log) subsystem. The case history is stored as a line-delimited file where each line represents a single event attributed to a specific user. User-controlled fields, including user names, event messages, and case titles used during case merging, were written to the history file without sanitizing embedded line separator characters (such as carriage return, line feed, vertical tab, form feed, and Unicode line/paragraph separators). An authenticated user with the ability to create or modify case history entries could embed line separators in these fields. Upon writing, the injected separators would split a single intended entry into multiple lines, each of which the history format interprets as a distinct event. This allows an attacker to forge additional history entries attributed to arbitrary users, corrupting the audit trail and undermining the integrity of case records. The vulnerability affects the case history append and merge operations. No authentication bypass or privilege escalation is required beyond normal case-management access. Affected versions: flowintel prior to 4.1.0.
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 08:42 UTC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:16 UTC
CWE
  • CWE-117 - Improper Output Neutralization for Logs
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 08:42
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1.patch eab799b16980…
Confidence
medium
Commit Subject Patch SHA-256
842a948507b8 fix: [security] Keep case history entries on a single line eab799b16980…
Fix summary

The fix introduces a sanitization helper that replaces all splitlines()-recognized line separator characters (CR, LF, VT, FF, file/group/record separators, NEL, Unicode line and paragraph separators) with a single space before any user-authored content is written to the case history file. The helper is applied to the author name, the event message, and the case title during merge operations, ensuring each history entry remains on exactly one line and cannot be split into forged entries.

Patch summary

In app/case/common_core.py, a compiled regex _HISTORY_LINE_BREAKS matching all splitlines() separators is added, along with a helper function _history_one_line() that substitutes matches with a space. In save_history(), the author name and message are passed through _history_one_line() before being written. In append_history_file(), each source line and the merging case title are also sanitized via _history_one_line() before being appended to the destination history file, and an explicit newline is written to terminate each merged entry.

CVSS rationale

AV:N: The application is a web-based case management tool; input is submitted over the network. AC:L: Exploitation requires only embedding a newline character in a name, message, or title field; no race conditions or special timing. AT:N: No special attack conditions are needed. PR:L: The attacker must be an authenticated user with access to create or modify case history entries. UI:N: No victim interaction is required beyond the attacker's own submission. VC:N: No confidentiality impact; the vulnerability does not expose data. VI:L: Integrity impact is limited to forging entries in a single case's history file; it does not compromise the broader system or other cases. VA:N: No availability impact. SC/SI/SA:N: The impact is confined to the application's own log file and does not cross a security scope boundary.

Weakness rationale
  • CWE-117 User-controlled data (names, messages, titles) is written to a structured log file without neutralizing line separator characters, allowing injection of additional log entries. This is the textbook definition of CWE-117.
Attack pattern rationale
  • CAPEC-66 CAPEC-66 describes the technique of injecting CRLF or line separator characters into structured output to create additional entries or split a single record into multiple. The underlying mechanism here is identical: an attacker embeds line separators in user input to split a single log entry into multiple forged entries. The context differs (log file rather than HTTP response), but no CAPEC pattern more precisely captures structured-output injection via line separators. This is the closest available match.
Assumptions to verify
  • PR:L assumes the attacker needs at least basic authenticated access to the case management interface to submit history entries; the patch does not specify whether unauthenticated access is possible.
  • VI:L assumes the integrity impact is limited to the case history file and does not extend to other system components or databases; the patch only modifies the file-writing path.
  • CAPEC-66 is the closest available pattern for structured-output injection via line separators; no CAPEC specifically covers log-file injection, so this mapping carries some uncertainty.
  • The affected version range (0 to <4.1.0) is taken from the explicit analyst context; the patch itself does not encode version boundaries.
  • The commit message references an internal finding identifier (F124 / G29) which is not publicly resolvable and is not included as a reference.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 7 11 medium 5
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-06 09:54 UTC Fix developed Corrective change authored (842a948507b87ee8c614ba2871ffecb3172d29e1): fix: [security] Keep case history entries on a single line https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1.patch
  2. 2026-11-20 08:41 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133627",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:16:19.260199Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:16:37.656Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/case/common_core.py"
          ],
          "product": "flowintel",
          "programFiles": [
            "app/case/common_core.py"
          ],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "flowintel contains a log injection vulnerability in its case history (audit log) subsystem. The case history is stored as a line-delimited file where each line represents a single event attributed to a specific user. User-controlled fields, including user names, event messages, and case titles used during case merging, were written to the history file without sanitizing embedded line separator characters (such as carriage return, line feed, vertical tab, form feed, and Unicode line/paragraph separators).\n\nAn authenticated user with the ability to create or modify case history entries could embed line separators in these fields. Upon writing, the injected separators would split a single intended entry into multiple lines, each of which the history format interprets as a distinct event. This allows an attacker to forge additional history entries attributed to arbitrary users, corrupting the audit trail and undermining the integrity of case records.\n\nThe vulnerability affects the case history append and merge operations. No authentication bypass or privilege escalation is required beyond normal case-management access.\n\nAffected versions: flowintel prior to 4.1.0."
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-66",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-66 HTTP Request/Response Splitting"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 5.3,
            "baseSeverity": "MEDIUM",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:L/VA:N/SC:N/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "LOW",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T08:42:41Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-117",
              "description": "CWE-117 Improper Output Neutralization for Logs",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T08:52:54.104Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The fix introduces a sanitization helper that replaces all splitlines()-recognized line separator characters (CR, LF, VT, FF, file/group/record separators, NEL, Unicode line and paragraph separators) with a single space before any user-authored content is written to the case history file. The helper is applied to the author name, the event message, and the case title during merge operations, ensuring each history entry remains on exactly one line and cannot be split into forged entries."
        }
      ],
      "title": "flowintel Case History Log Injection via Line Separator Characters in User-Authored Fields",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "PR:L assumes the attacker needs at least basic authenticated access to the case management interface to submit history entries; the patch does not specify whether unauthenticated access is possible.",
                  "VI:L assumes the integrity impact is limited to the case history file and does not extend to other system components or databases; the patch only modifies the file-writing path.",
                  "CAPEC-66 is the closest available pattern for structured-output injection via line separators; no CAPEC specifically covers log-file injection, so this mapping carries some uncertainty.",
                  "The affected version range (0 to \u003c4.1.0) is taken from the explicit analyst context; the patch itself does not encode version boundaries.",
                  "The commit message references an internal finding identifier (F124 / G29) which is not publicly resolvable and is not included as a reference."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-66",
                    "rationale": "CAPEC-66 describes the technique of injecting CRLF or line separator characters into structured output to create additional entries or split a single record into multiple. The underlying mechanism here is identical: an attacker embeds line separators in user input to split a single log entry into multiple forged entries. The context differs (log file rather than HTTP response), but no CAPEC pattern more precisely captures structured-output injection via line separators. This is the closest available match."
                  }
                ],
                "commit": "842a948507b87ee8c614ba2871ffecb3172d29e1",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N: The application is a web-based case management tool; input is submitted over the network. AC:L: Exploitation requires only embedding a newline character in a name, message, or title field; no race conditions or special timing. AT:N: No special attack conditions are needed. PR:L: The attacker must be an authenticated user with access to create or modify case history entries. UI:N: No victim interaction is required beyond the attacker\u0027s own submission. VC:N: No confidentiality impact; the vulnerability does not expose data. VI:L: Integrity impact is limited to forging entries in a single case\u0027s history file; it does not compromise the broader system or other cases. VA:N: No availability impact. SC/SI/SA:N: The impact is confined to the application\u0027s own log file and does not cross a security scope boundary.",
                "fixSummary": "The fix introduces a sanitization helper that replaces all splitlines()-recognized line separator characters (CR, LF, VT, FF, file/group/record separators, NEL, Unicode line and paragraph separators) with a single space before any user-authored content is written to the case history file. The helper is applied to the author name, the event message, and the case title during merge operations, ensuring each history entry remains on exactly one line and cannot be split into forged entries.",
                "generatedAt": "2026-11-20T08:42:41.961994Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 5,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 7
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "eab799b16980a28458755cce56ea9897af43b9a630aec94b557deb9c8f373ab4",
                "patchSummary": "In app/case/common_core.py, a compiled regex _HISTORY_LINE_BREAKS matching all splitlines() separators is added, along with a helper function _history_one_line() that substitutes matches with a space. In save_history(), the author name and message are passed through _history_one_line() before being written. In append_history_file(), each source line and the merging case title are also sanitized via _history_one_line() before being appended to the destination history file, and an explicit newline is written to terminate each merged entry.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "842a948507b87ee8c614ba2871ffecb3172d29e1",
                    "date": "Fri, 6 Nov 2026 10:54:12 +0100",
                    "patchSha256": "eab799b16980a28458755cce56ea9897af43b9a630aec94b557deb9c8f373ab4",
                    "retrievedAt": "2026-11-20T08:41:59.334828Z",
                    "source": "https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1.patch",
                    "subject": "fix: [security] Keep case history entries on a single line"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T08:42:41Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Keep case history entries on a single line",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-117",
                    "rationale": "User-controlled data (names, messages, titles) is written to a structured log file without neutralizing line separator characters, allowing injection of additional log entries. This is the textbook definition of CWE-117."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (842a948507b87ee8c614ba2871ffecb3172d29e1): fix: [security] Keep case history entries on a single line",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1.patch"
                    ],
                    "timestamp": "2026-11-06T09:54:12Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1.patch"
                    ],
                    "timestamp": "2026-11-20T08:41:59.334828Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20212"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133627",
    "datePublished": "2026-11-20T08:52:54.104Z",
    "dateReserved": "2026-11-20T08:52:49.070Z",
    "dateUpdated": "2026-11-20T18:16:37.656Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133626 (GCVE-0-2026-133626)

Vulnerability from cvelistv5 – Published: 2026-11-20 08:52 – Updated: 2026-11-20 18:15
VLAI
Title
FlowIntel: Resource Exhaustion via Sparse CSV/JSON Table Expansion During File-to-Note Conversion
Summary
FlowIntel contains a resource exhaustion vulnerability in its file-to-note conversion feature. When a user converts an attached CSV or JSON file into a case note, the application renders the data as a Markdown table by padding every row to the maximum column count. A crafted file with a small number of filled cells but a very large number of columns (or keys) causes the application to allocate a quadratic number of padded table cells, potentially exhausting a worker process in memory or CPU. The same issue affects nested JSON values rendered inside table cells, where pretty-printed indentation further inflates the output size. Preconditions: - An authenticated user with access to a case in FlowIntel - Ability to attach and convert a CSV or JSON file to a note Impact: - Denial of service: exhaustion of a worker process, potentially disrupting service for other users - No confidentiality or integrity impact Affected versions: FlowIntel prior to 4.1.0
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 08:43 UTC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:15 UTC
CWE
  • CWE-400 - Uncontrolled Resource Consumption
  • CWE-770 - Allocation of Resources Without Limits or Throttling
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 08:43
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6.patch 1aa2d65a4e08…
Confidence
high
Commit Subject Patch SHA-256
90b822a0fca3 fix: [security] Bound CSV and JSON expansion when converting 1aa2d65a4e08…
Fix summary

The fix introduces bounds on table expansion: a maximum padded-cell count (50,000) and a maximum padding ratio (4x filled cells). When a CSV or JSON dataset would exceed these limits, the original content is preserved in a safely fenced code block instead of being rendered as a table. Additionally, nested JSON values in table cells are now serialized in compact form (no indentation), and the JSON fallback path uses the same safe fencing helper, preventing both quadratic cell expansion and indentation-based output inflation.

Patch summary

In app/utils/file_converter.py: added constants MAX_PADDED_TABLE_CELLS (50000) and MAX_PADDING_RATIO (4); added helper _table_too_sparse() to detect oversized sparse tables; added helper _fenced_content() that generates a code fence long enough to prevent fence-injection from user content; modified convert_csv_to_note() to short-circuit to fenced output when the table is too sparse; modified convert_json_to_note() to apply the same sparse check for arrays of objects; changed json.dumps(value, indent=2) to json.dumps(value) for nested values in table cells; replaced the JSON fallback json.dumps(data, indent=2) with _fenced_content(content, "json"). In doc/user-manual.md: added a note documenting the new sparse-data behavior.

CVSS rationale

AV:N - the file conversion is accessible over the network via the web application. AC:L - crafting a sparse CSV/JSON file is straightforward; no race conditions or special timing required. AT:N - no prior tampering with system data needed. PR:L - requires an authenticated user with access to a case to attach and convert a file. UI:N - the conversion is triggered by the attacker's own action (uploading and converting the file); no victim interaction needed. VC:N, VI:N - no confidentiality or integrity impact. VA:H - worker exhaustion can deny service to other users of the FlowIntel instance. SC:N, SI:N, SA:N - no secondary component impact is evidenced.

Weakness rationale
  • CWE-400 The application does not bound the expansion of sparse CSV/JSON data into padded table cells, allowing a small input to trigger quadratic memory/CPU allocation and exhaust a worker. This is a classic uncontrolled resource consumption issue.
  • CWE-770 The table rendering path allocates cells proportional to rows x max_cols without any upper bound, and JSON pretty-printing further inflates output. No limit or throttle was applied before the fix.
Attack pattern rationale
  • CAPEC-10 The attacker supplies a crafted CSV or JSON file that, when processed by the conversion feature, causes the application to allocate an excessive number of table cells, exhausting worker resources. This is a direct match to application resource exhaustion via crafted input.
  • CAPEC-124 The underlying mechanism is quadratic expansion: a file with N columns and M rows produces N*M padded cells. The attacker exploits the algorithmic complexity of the table-padding logic. This is a secondary, more specific mapping; CAPEC-10 is the primary match.
Assumptions to verify
  • PR:L assumes the file-to-note conversion feature requires an authenticated user with case access; the patch does not explicitly state authentication requirements, but FlowIntel is a case-management tool where case operations are typically authenticated.
  • VA:H assumes that exhausting a single worker has a meaningful availability impact on the service; if the deployment runs many redundant workers behind a load balancer, the practical impact could be lower.
  • CAPEC-124 is included as a secondary mapping; the primary mechanism is unbounded allocation (CAPEC-10), and the quadratic complexity is the underlying cause rather than a separate attack pattern.
  • The affected version range is taken from the explicit context (affected_version 0, fixed_version 4.1.0); the patch itself does not encode version boundaries.
  • The _fenced_content helper also mitigates a potential Markdown fence-injection issue (a crafted backtick sequence in the file content could break out of the code block), but the primary vulnerability addressed is resource exhaustion.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 8 11 high 5
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-06 09:53 UTC Fix developed Corrective change authored (90b822a0fca35d33688730772b964b478c7fcde6): fix: [security] Bound CSV and JSON expansion when converting https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6.patch
  2. 2026-11-20 08:42 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133626",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:15:29.296559Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:15:48.121Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/utils/file_converter.py"
          ],
          "product": "flowintel",
          "programFiles": [
            "app/utils/file_converter.py"
          ],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "FlowIntel contains a resource exhaustion vulnerability in its file-to-note conversion feature. When a user converts an attached CSV or JSON file into a case note, the application renders the data as a Markdown table by padding every row to the maximum column count. A crafted file with a small number of filled cells but a very large number of columns (or keys) causes the application to allocate a quadratic number of padded table cells, potentially exhausting a worker process in memory or CPU.\n\nThe same issue affects nested JSON values rendered inside table cells, where pretty-printed indentation further inflates the output size.\n\nPreconditions:\n- An authenticated user with access to a case in FlowIntel\n- Ability to attach and convert a CSV or JSON file to a note\n\nImpact:\n- Denial of service: exhaustion of a worker process, potentially disrupting service for other users\n- No confidentiality or integrity impact\n\nAffected versions: FlowIntel prior to 4.1.0"
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-10",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-10 Application Resource Exhaustion"
            }
          ]
        },
        {
          "capecId": "CAPEC-124",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-124 Resource Exhaustion via Algorithmic Complexity"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 7.1,
            "baseSeverity": "HIGH",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "HIGH",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "NONE",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T08:43:07Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-400",
              "description": "CWE-400 Uncontrolled Resource Consumption",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-770",
              "description": "CWE-770 Allocation of Resources Without Limits or Throttling",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T08:52:28.073Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The fix introduces bounds on table expansion: a maximum padded-cell count (50,000) and a maximum padding ratio (4x filled cells). When a CSV or JSON dataset would exceed these limits, the original content is preserved in a safely fenced code block instead of being rendered as a table. Additionally, nested JSON values in table cells are now serialized in compact form (no indentation), and the JSON fallback path uses the same safe fencing helper, preventing both quadratic cell expansion and indentation-based output inflation."
        }
      ],
      "title": "FlowIntel: Resource Exhaustion via Sparse CSV/JSON Table Expansion During File-to-Note Conversion",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "PR:L assumes the file-to-note conversion feature requires an authenticated user with case access; the patch does not explicitly state authentication requirements, but FlowIntel is a case-management tool where case operations are typically authenticated.",
                  "VA:H assumes that exhausting a single worker has a meaningful availability impact on the service; if the deployment runs many redundant workers behind a load balancer, the practical impact could be lower.",
                  "CAPEC-124 is included as a secondary mapping; the primary mechanism is unbounded allocation (CAPEC-10), and the quadratic complexity is the underlying cause rather than a separate attack pattern.",
                  "The affected version range is taken from the explicit context (affected_version 0, fixed_version 4.1.0); the patch itself does not encode version boundaries.",
                  "The _fenced_content helper also mitigates a potential Markdown fence-injection issue (a crafted backtick sequence in the file content could break out of the code block), but the primary vulnerability addressed is resource exhaustion."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-10",
                    "rationale": "The attacker supplies a crafted CSV or JSON file that, when processed by the conversion feature, causes the application to allocate an excessive number of table cells, exhausting worker resources. This is a direct match to application resource exhaustion via crafted input."
                  },
                  {
                    "capecId": "CAPEC-124",
                    "rationale": "The underlying mechanism is quadratic expansion: a file with N columns and M rows produces N*M padded cells. The attacker exploits the algorithmic complexity of the table-padding logic. This is a secondary, more specific mapping; CAPEC-10 is the primary match."
                  }
                ],
                "commit": "90b822a0fca35d33688730772b964b478c7fcde6",
                "confidence": "high",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N - the file conversion is accessible over the network via the web application. AC:L - crafting a sparse CSV/JSON file is straightforward; no race conditions or special timing required. AT:N - no prior tampering with system data needed. PR:L - requires an authenticated user with access to a case to attach and convert a file. UI:N - the conversion is triggered by the attacker\u0027s own action (uploading and converting the file); no victim interaction needed. VC:N, VI:N - no confidentiality or integrity impact. VA:H - worker exhaustion can deny service to other users of the FlowIntel instance. SC:N, SI:N, SA:N - no secondary component impact is evidenced.",
                "fixSummary": "The fix introduces bounds on table expansion: a maximum padded-cell count (50,000) and a maximum padding ratio (4x filled cells). When a CSV or JSON dataset would exceed these limits, the original content is preserved in a safely fenced code block instead of being rendered as a table. Additionally, nested JSON values in table cells are now serialized in compact form (no indentation), and the JSON fallback path uses the same safe fencing helper, preventing both quadratic cell expansion and indentation-based output inflation.",
                "generatedAt": "2026-11-20T08:43:07.422130Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 5,
                      "confidence": "high",
                      "model": "qwen3.8:27b",
                      "score": 8
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "1aa2d65a4e08d2c447aea812aa20e4190cf65a5be1e6a751afaae121545e51f4",
                "patchSummary": "In app/utils/file_converter.py: added constants MAX_PADDED_TABLE_CELLS (50000) and MAX_PADDING_RATIO (4); added helper _table_too_sparse() to detect oversized sparse tables; added helper _fenced_content() that generates a code fence long enough to prevent fence-injection from user content; modified convert_csv_to_note() to short-circuit to fenced output when the table is too sparse; modified convert_json_to_note() to apply the same sparse check for arrays of objects; changed json.dumps(value, indent=2) to json.dumps(value) for nested values in table cells; replaced the JSON fallback json.dumps(data, indent=2) with _fenced_content(content, \"json\"). In doc/user-manual.md: added a note documenting the new sparse-data behavior.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "90b822a0fca35d33688730772b964b478c7fcde6",
                    "date": "Fri, 6 Nov 2026 10:53:20 +0100",
                    "patchSha256": "1aa2d65a4e08d2c447aea812aa20e4190cf65a5be1e6a751afaae121545e51f4",
                    "retrievedAt": "2026-11-20T08:42:42.445428Z",
                    "source": "https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6.patch",
                    "subject": "fix: [security] Bound CSV and JSON expansion when converting"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T08:43:07Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Bound CSV and JSON expansion when converting",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-400",
                    "rationale": "The application does not bound the expansion of sparse CSV/JSON data into padded table cells, allowing a small input to trigger quadratic memory/CPU allocation and exhaust a worker. This is a classic uncontrolled resource consumption issue."
                  },
                  {
                    "cweId": "CWE-770",
                    "rationale": "The table rendering path allocates cells proportional to rows x max_cols without any upper bound, and JSON pretty-printing further inflates output. No limit or throttle was applied before the fix."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (90b822a0fca35d33688730772b964b478c7fcde6): fix: [security] Bound CSV and JSON expansion when converting",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6.patch"
                    ],
                    "timestamp": "2026-11-06T09:53:20Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6.patch"
                    ],
                    "timestamp": "2026-11-20T08:42:42.445428Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20366"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133626",
    "datePublished": "2026-11-20T08:52:28.073Z",
    "dateReserved": "2026-11-20T08:52:20.420Z",
    "dateUpdated": "2026-11-20T18:15:48.121Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}
displaying 1 - 10 vulnerabilities in total 362

GCVE-1-2026-20423 (CVE-2026-133855)

Vulnerability from gna-1 – Published: 2026-11-20 14:44 – Updated: 2026-11-20 14:44
VLAI
Title
flowintel: Unauthorized Disclosure of Linked Case References in Case History
Summary
flowintel contains an information disclosure vulnerability in its case history feature. When a user views the history of a case they are authorized to access, the history output includes references to other cases (linked, unlinked, or forked) without verifying whether the requesting user has permission to view those referenced cases. This exposes the identifiers and titles of cases that the user should not be able to discover. The vulnerability affects multiple output paths: - Browser-based case history view - API history endpoint - Plain-text history download - Markdown history download - Report timeline generation An authenticated user with access to any single case can enumerate the existence and identifiers of other cases by reading the history of a case they can access. This bypasses the intended per-case access control model and may reveal the existence of sensitive investigations, their identifiers, and associated metadata to unauthorized users. The raw audit storage is not modified; the issue is solely in the presentation layer failing to filter references to inaccessible cases before returning history data to the user. Affected versions: flowintel prior to 4.1.0.
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CNA (v2.0.3)
Decision recorded 2026-11-20 14:43 UTC
CWE
  • CWE-200 - Exposure of Sensitive Information to an Unauthorized Actor
  • CWE-862 - Missing Authorization
Assigner
CIRCL GNA GNA-1
GNA scorecard A 92/100 over 319 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 14:43
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/c44025ea68109996e3b1a0f01f6333caba190809 407e73d34f15…
Confidence
medium
Commit Subject Patch SHA-256
c44025ea6810 fix: [security] Filter linked case references from case 407e73d34f15…
Fix summary

A new visibility-aware history retrieval function was introduced that parses each history record for case references (link, unlink, fork events) and checks the requesting user's access to every referenced case before including that record in the output. Records referencing inaccessible cases are omitted from the response. The filtering is applied consistently across all history output paths (browser view, API, text download, Markdown download, and report timelines). Administrative users retain full unfiltered access. The raw audit log storage remains unchanged to preserve forensic integrity.

Patch summary

Added a new function get_visible_history() in common_core.py that: (1) verifies the user can view the current case, (2) groups raw history lines into logical records using timestamp-based event boundaries, (3) extracts referenced case IDs from link/unlink/fork messages using regex patterns, (4) checks can_view_case() for each referenced case, and (5) excludes records referencing inaccessible cases. Updated download_history(), download_history_md(), the browser history endpoint, the API history endpoint, and the report timeline generator to call get_visible_history() instead of the raw get_history(). Added helper functions _history_case_references() and _history_record_complete() to handle legacy multiline link records and ambiguous formats. Changed download_history() to serve content via BytesIO instead of reading a file path directly.

CVSS rationale

AV:N - The vulnerability is exploitable over the network via the web application. AC:L - Exploitation requires only viewing the history of any accessible case; no special conditions or race conditions are needed. AT:N - No prior attack or manipulation of the target is required. PR:L - The attacker must be an authenticated user with access to at least one case. UI:N - No user interaction beyond normal application use is required. VC:L - The vulnerable component (case history endpoint) leaks limited information (case IDs and titles of referenced cases). VI:N - No integrity impact. VA:N - No availability impact. SC:L - Subsequent components (other cases) have their existence and identifiers exposed to unauthorized users. SI:N - No integrity impact on other systems. SA:N - No safety impact.

Weakness rationale
  • CWE-200 The case history output exposes identifiers and titles of other cases to users who do not have authorization to view those cases. The sensitive data (case existence, IDs, titles) is disclosed through a legitimate endpoint without proper per-reference authorization checks.
  • CWE-862 The history retrieval functions did not perform authorization checks on the referenced cases embedded within the history records. The primary case access was verified, but the secondary cases referenced in link/unlink/fork events were not checked before inclusion in the response.
Attack pattern rationale
  • CAPEC-128 The attacker leverages a legitimate application endpoint (case history) to discover metadata about other resources (case identifiers, titles, and relationships) that they are not authorized to access. This is the closest CAPEC pattern to the observed information disclosure through a functional endpoint. The mapping is approximate because CAPEC-128 typically refers to version/configuration discovery, whereas here the disclosed metadata is application data (case references) rather than system metadata. No more specific CAPEC pattern for cross-object information disclosure via a legitimate read endpoint exists in the CAPEC catalog.
Assumptions to verify
  • The affected version range is assumed to be all versions prior to 4.1.0 based on the explicit context providing affected_version '0' and fixed_version '4.1.0'. The exact earliest affected version is not confirmed by the patch itself.
  • The CAPEC-128 mapping is the closest available pattern; no CAPEC specifically covers cross-object information disclosure through a legitimate read endpoint. The mapping is approximate.
  • CVSS PR:L assumes the attacker needs at least one valid authenticated session with access to any case. If the application allows unauthenticated history access in some deployment, PR would be N.
  • The patch commit message references 'Tool: Codex' which is noted as the AI tool used for implementation but is not credited as a vulnerability finder or reporter.
  • The vulnerability requires the attacker to already have access to at least one case; it does not allow access to cases with no prior relationship to the attacker's accessible cases beyond those referenced in history.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 7 11 medium 5
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-17 12:50 UTC Fix developed Corrective change authored (c44025ea68109996e3b1a0f01f6333caba190809): fix: [security] Filter linked case references from case https://github.com/flowintel/flowintel/commit/c44025ea68109996e3b1a0f01f6333caba190809
  2. 2026-11-20 14:42 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/c44025ea68109996e3b1a0f01f6333caba190809

{
  "containers": {
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/case/CaseCore.py",
            "app/case/case.py",
            "app/case/case_api.py",
            "app/case/case_notes.py",
            "app/case/common_core.py"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "flowintel contains an information disclosure vulnerability in its case history feature. When a user views the history of a case they are authorized to access, the history output includes references to other cases (linked, unlinked, or forked) without verifying whether the requesting user has permission to view those referenced cases. This exposes the identifiers and titles of cases that the user should not be able to discover.\n\nThe vulnerability affects multiple output paths:\n- Browser-based case history view\n- API history endpoint\n- Plain-text history download\n- Markdown history download\n- Report timeline generation\n\nAn authenticated user with access to any single case can enumerate the existence and identifiers of other cases by reading the history of a case they can access. This bypasses the intended per-case access control model and may reveal the existence of sensitive investigations, their identifiers, and associated metadata to unauthorized users.\n\nThe raw audit storage is not modified; the issue is solely in the presentation layer failing to filter references to inaccessible cases before returning history data to the user.\n\nAffected versions: flowintel prior to 4.1.0."
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-128",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-128 Application Metadata Discovery"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 5.3,
            "baseSeverity": "MEDIUM",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "LOW",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:N/VA:N/SC:L/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "LOW",
            "vulnIntegrityImpact": "NONE",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T14:43:12Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-200",
              "description": "CWE-200 Exposure of Sensitive Information to an Unauthorized Actor",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-862",
              "description": "CWE-862 Missing Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "orgId": "00000000-0000-4000-8000-000000000000"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/c44025ea68109996e3b1a0f01f6333caba190809"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "A new visibility-aware history retrieval function was introduced that parses each history record for case references (link, unlink, fork events) and checks the requesting user\u0027s access to every referenced case before including that record in the output. Records referencing inaccessible cases are omitted from the response. The filtering is applied consistently across all history output paths (browser view, API, text download, Markdown download, and report timelines). Administrative users retain full unfiltered access. The raw audit log storage remains unchanged to preserve forensic integrity."
        }
      ],
      "title": "flowintel: Unauthorized Disclosure of Linked Case References in Case History",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected version range is assumed to be all versions prior to 4.1.0 based on the explicit context providing affected_version \u00270\u0027 and fixed_version \u00274.1.0\u0027. The exact earliest affected version is not confirmed by the patch itself.",
                  "The CAPEC-128 mapping is the closest available pattern; no CAPEC specifically covers cross-object information disclosure through a legitimate read endpoint. The mapping is approximate.",
                  "CVSS PR:L assumes the attacker needs at least one valid authenticated session with access to any case. If the application allows unauthenticated history access in some deployment, PR would be N.",
                  "The patch commit message references \u0027Tool: Codex\u0027 which is noted as the AI tool used for implementation but is not credited as a vulnerability finder or reporter.",
                  "The vulnerability requires the attacker to already have access to at least one case; it does not allow access to cases with no prior relationship to the attacker\u0027s accessible cases beyond those referenced in history."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-128",
                    "rationale": "The attacker leverages a legitimate application endpoint (case history) to discover metadata about other resources (case identifiers, titles, and relationships) that they are not authorized to access. This is the closest CAPEC pattern to the observed information disclosure through a functional endpoint. The mapping is approximate because CAPEC-128 typically refers to version/configuration discovery, whereas here the disclosed metadata is application data (case references) rather than system metadata. No more specific CAPEC pattern for cross-object information disclosure via a legitimate read endpoint exists in the CAPEC catalog."
                  }
                ],
                "commit": "c44025ea68109996e3b1a0f01f6333caba190809",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N - The vulnerability is exploitable over the network via the web application. AC:L - Exploitation requires only viewing the history of any accessible case; no special conditions or race conditions are needed. AT:N - No prior attack or manipulation of the target is required. PR:L - The attacker must be an authenticated user with access to at least one case. UI:N - No user interaction beyond normal application use is required. VC:L - The vulnerable component (case history endpoint) leaks limited information (case IDs and titles of referenced cases). VI:N - No integrity impact. VA:N - No availability impact. SC:L - Subsequent components (other cases) have their existence and identifiers exposed to unauthorized users. SI:N - No integrity impact on other systems. SA:N - No safety impact.",
                "fixSummary": "A new visibility-aware history retrieval function was introduced that parses each history record for case references (link, unlink, fork events) and checks the requesting user\u0027s access to every referenced case before including that record in the output. Records referencing inaccessible cases are omitted from the response. The filtering is applied consistently across all history output paths (browser view, API, text download, Markdown download, and report timelines). Administrative users retain full unfiltered access. The raw audit log storage remains unchanged to preserve forensic integrity.",
                "generatedAt": "2026-11-20T14:43:12.693948Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 5,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 7
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "407e73d34f154f6febfa70e71f7a7c5816cae0d73d6b25d0e6f4a411304336cd",
                "patchSummary": "Added a new function get_visible_history() in common_core.py that: (1) verifies the user can view the current case, (2) groups raw history lines into logical records using timestamp-based event boundaries, (3) extracts referenced case IDs from link/unlink/fork messages using regex patterns, (4) checks can_view_case() for each referenced case, and (5) excludes records referencing inaccessible cases. Updated download_history(), download_history_md(), the browser history endpoint, the API history endpoint, and the report timeline generator to call get_visible_history() instead of the raw get_history(). Added helper functions _history_case_references() and _history_record_complete() to handle legacy multiline link records and ambiguous formats. Changed download_history() to serve content via BytesIO instead of reading a file path directly.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "c44025ea68109996e3b1a0f01f6333caba190809",
                    "date": "Tue, 17 Nov 2026 13:50:29 +0100",
                    "patchSha256": "407e73d34f154f6febfa70e71f7a7c5816cae0d73d6b25d0e6f4a411304336cd",
                    "retrievedAt": "2026-11-20T14:42:40.710956Z",
                    "source": "https://github.com/flowintel/flowintel/commit/c44025ea68109996e3b1a0f01f6333caba190809",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/c44025ea68109996e3b1a0f01f6333caba190809",
                    "subject": "fix: [security] Filter linked case references from case"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/c44025ea68109996e3b1a0f01f6333caba190809",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T14:43:12Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Filter linked case references from case",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-200",
                    "rationale": "The case history output exposes identifiers and titles of other cases to users who do not have authorization to view those cases. The sensitive data (case existence, IDs, titles) is disclosed through a legitimate endpoint without proper per-reference authorization checks."
                  },
                  {
                    "cweId": "CWE-862",
                    "rationale": "The history retrieval functions did not perform authorization checks on the referenced cases embedded within the history records. The primary case access was verified, but the secondary cases referenced in link/unlink/fork events were not checked before inclusion in the response."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (c44025ea68109996e3b1a0f01f6333caba190809): fix: [security] Filter linked case references from case",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/c44025ea68109996e3b1a0f01f6333caba190809"
                    ],
                    "timestamp": "2026-11-17T12:50:29Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/c44025ea68109996e3b1a0f01f6333caba190809"
                    ],
                    "timestamp": "2026-11-20T14:42:40.710956Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20423"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "00000000-0000-4000-8000-000000000000",
    "cveId": "CVE-2026-133855",
    "datePublished": "2026-11-20T14:44:18.699005Z",
    "dateReserved": "2026-11-20T14:44:53.894Z",
    "dateUpdated": "2026-11-20T14:44:54.833278Z",
    "state": "PUBLISHED",
    "vulnId": "GCVE-1-2026-20423"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

GCVE-1-2026-20422 (CVE-2026-133854)

Vulnerability from gna-1 – Published: 2026-11-20 14:44 – Updated: 2026-11-20 14:44
VLAI
Title
flowintel: Information Disclosure of Classification Usage via Missing Access Control
Summary
flowintel contains a vulnerability in the case and task classification usage endpoints that allows authenticated users to view classification data associated with cases they do not have permission to access. The affected endpoints, which provide lists of used taxonomies, tags, galaxies, clusters, and custom tags, did not properly enforce object-level access controls. While the endpoints required authentication, they failed to filter the returned data based on the requesting user's visibility permissions for the underlying cases. This allowed any authenticated user to enumerate classification usage across all cases, including private or restricted cases belonging to other organizations or users. This issue could lead to the disclosure of sensitive information, such as the existence of specific cases, the types of threats or indicators associated with them, and the relationships between different classification elements, potentially aiding an attacker in reconnaissance or targeting. Affected versions: < 4.1.0
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CNA (v2.0.3)
Decision recorded 2026-11-20 14:42 UTC
CWE
Assigner
CIRCL GNA GNA-1
GNA scorecard A 92/100 over 319 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 14:42
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/fa7216e760413f9a3feab9eca710b04cc8f2f630 58589e54911e…
Confidence
medium
Commit Subject Patch SHA-256
fa7216e76041 fix: [security] Scope classification usage to visible cases 58589e54911e…
Fix summary

The vulnerability is remediated by enforcing object-level access controls on the classification usage endpoints. The fix ensures that the authenticated user's identity is passed to the data retrieval functions, which then filter the results to include only classifications associated with cases that the user is authorized to view. This prevents the disclosure of classification data from private or restricted cases to unauthorized users.

Patch summary

The patch modifies the case and task classification usage endpoints to pass the current authenticated user to the underlying data retrieval functions. A new helper function, _classification_usage_ids, is introduced to filter classification usage based on the user's permissions. This function joins the classification link models with the Case model and applies filters to ensure that only classifications associated with visible cases (public, member, or administrator access) are returned. The existing helper functions for retrieving used taxonomies, tags, galaxies, clusters, and custom tags are updated to use this new filtering logic.

CVSS rationale

The vulnerability is exploitable over the network (AV:N) with low complexity (AC:L) and no attack requirements (AT:N). It requires low privileges (PR:L) as any authenticated user can exploit it. No user interaction is required (UI:N). The impact is primarily on the confidentiality of the vulnerable system (VI:H) as it allows unauthorized access to sensitive classification data. There is no impact on integrity or availability (VC:N, VA:N, SC:N, SI:N, SA:N).

Weakness rationale
  • CWE-862 The endpoints lacked proper authorization checks to ensure that users could only view classification data associated with cases they had permission to access. This is a classic missing authorization vulnerability where the system failed to verify that the user had the necessary permissions to access the specific objects (cases) referenced by the classification data.
Attack pattern rationale
  • CAPEC-126 The vulnerability allows an attacker to access classification data associated with cases they do not have permission to view by simply querying the endpoints. This aligns with the attack pattern of accessing functionality or data that is not properly restricted by identifiers or access controls. The uncertainty lies in the fact that the attacker does not need to guess specific IDs, but rather the lack of filtering exposes data across all cases, which is a broader form of improper restriction.
Assumptions to verify
  • The CAPEC mapping is uncertain as the vulnerability is more about missing filtering than accessing specific IDs, but CAPEC-126 is the closest plausible match.
  • The CVSS vector assumes that the impact is primarily on confidentiality and that the vulnerability does not lead to integrity or availability impacts.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 10 11 medium 2
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-17 12:03 UTC Fix developed Corrective change authored (fa7216e760413f9a3feab9eca710b04cc8f2f630): fix: [security] Scope classification usage to visible cases https://github.com/flowintel/flowintel/commit/fa7216e760413f9a3feab9eca710b04cc8f2f630
  2. 2026-11-20 14:41 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/fa7216e760413f9a3feab9eca710b04cc8f2f630

{
  "containers": {
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/case/case.py",
            "app/case/common_core.py"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "flowintel contains a vulnerability in the case and task classification usage endpoints that allows authenticated users to view classification data associated with cases they do not have permission to access.\n\nThe affected endpoints, which provide lists of used taxonomies, tags, galaxies, clusters, and custom tags, did not properly enforce object-level access controls. While the endpoints required authentication, they failed to filter the returned data based on the requesting user\u0027s visibility permissions for the underlying cases. This allowed any authenticated user to enumerate classification usage across all cases, including private or restricted cases belonging to other organizations or users.\n\nThis issue could lead to the disclosure of sensitive information, such as the existence of specific cases, the types of threats or indicators associated with them, and the relationships between different classification elements, potentially aiding an attacker in reconnaissance or targeting.\n\nAffected versions: \u003c 4.1.0"
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-126",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-126 Accessing Functionality Not Properly Restricted by ID"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 7.1,
            "baseSeverity": "HIGH",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:H/VA:N/SC:N/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "HIGH",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T14:42:40Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-862",
              "description": "CWE-862 Missing Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "orgId": "00000000-0000-4000-8000-000000000000"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/fa7216e760413f9a3feab9eca710b04cc8f2f630"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The vulnerability is remediated by enforcing object-level access controls on the classification usage endpoints. The fix ensures that the authenticated user\u0027s identity is passed to the data retrieval functions, which then filter the results to include only classifications associated with cases that the user is authorized to view. This prevents the disclosure of classification data from private or restricted cases to unauthorized users."
        }
      ],
      "title": "flowintel: Information Disclosure of Classification Usage via Missing Access Control",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The CAPEC mapping is uncertain as the vulnerability is more about missing filtering than accessing specific IDs, but CAPEC-126 is the closest plausible match.",
                  "The CVSS vector assumes that the impact is primarily on confidentiality and that the vulnerability does not lead to integrity or availability impacts."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-126",
                    "rationale": "The vulnerability allows an attacker to access classification data associated with cases they do not have permission to view by simply querying the endpoints. This aligns with the attack pattern of accessing functionality or data that is not properly restricted by identifiers or access controls. The uncertainty lies in the fact that the attacker does not need to guess specific IDs, but rather the lack of filtering exposes data across all cases, which is a broader form of improper restriction."
                  }
                ],
                "commit": "fa7216e760413f9a3feab9eca710b04cc8f2f630",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "The vulnerability is exploitable over the network (AV:N) with low complexity (AC:L) and no attack requirements (AT:N). It requires low privileges (PR:L) as any authenticated user can exploit it. No user interaction is required (UI:N). The impact is primarily on the confidentiality of the vulnerable system (VI:H) as it allows unauthorized access to sensitive classification data. There is no impact on integrity or availability (VC:N, VA:N, SC:N, SI:N, SA:N).",
                "fixSummary": "The vulnerability is remediated by enforcing object-level access controls on the classification usage endpoints. The fix ensures that the authenticated user\u0027s identity is passed to the data retrieval functions, which then filter the results to include only classifications associated with cases that the user is authorized to view. This prevents the disclosure of classification data from private or restricted cases to unauthorized users.",
                "generatedAt": "2026-11-20T14:42:40.009857Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 2,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 10
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "58589e54911e25f697c4143266f668e3726b793dddaddd6aa8d890cc059bb70d",
                "patchSummary": "The patch modifies the case and task classification usage endpoints to pass the current authenticated user to the underlying data retrieval functions. A new helper function, _classification_usage_ids, is introduced to filter classification usage based on the user\u0027s permissions. This function joins the classification link models with the Case model and applies filters to ensure that only classifications associated with visible cases (public, member, or administrator access) are returned. The existing helper functions for retrieving used taxonomies, tags, galaxies, clusters, and custom tags are updated to use this new filtering logic.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "fa7216e760413f9a3feab9eca710b04cc8f2f630",
                    "date": "Tue, 17 Nov 2026 13:03:32 +0100",
                    "patchSha256": "58589e54911e25f697c4143266f668e3726b793dddaddd6aa8d890cc059bb70d",
                    "retrievedAt": "2026-11-20T14:41:59.835058Z",
                    "source": "https://github.com/flowintel/flowintel/commit/fa7216e760413f9a3feab9eca710b04cc8f2f630",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/fa7216e760413f9a3feab9eca710b04cc8f2f630",
                    "subject": "fix: [security] Scope classification usage to visible cases"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/fa7216e760413f9a3feab9eca710b04cc8f2f630",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T14:42:40Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Scope classification usage to visible cases",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-862",
                    "rationale": "The endpoints lacked proper authorization checks to ensure that users could only view classification data associated with cases they had permission to access. This is a classic missing authorization vulnerability where the system failed to verify that the user had the necessary permissions to access the specific objects (cases) referenced by the classification data."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (fa7216e760413f9a3feab9eca710b04cc8f2f630): fix: [security] Scope classification usage to visible cases",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/fa7216e760413f9a3feab9eca710b04cc8f2f630"
                    ],
                    "timestamp": "2026-11-17T12:03:32Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/fa7216e760413f9a3feab9eca710b04cc8f2f630"
                    ],
                    "timestamp": "2026-11-20T14:41:59.835058Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20422"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "00000000-0000-4000-8000-000000000000",
    "cveId": "CVE-2026-133854",
    "datePublished": "2026-11-20T14:44:16.707290Z",
    "dateReserved": "2026-11-20T14:44:41.420Z",
    "dateUpdated": "2026-11-20T14:44:41.577159Z",
    "state": "PUBLISHED",
    "vulnId": "GCVE-1-2026-20422"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

GCVE-1-2026-20421 (CVE-2026-133853)

Vulnerability from gna-1 – Published: 2026-11-20 14:44 – Updated: 2026-11-20 14:44
VLAI
Title
FlowIntel Chatbot Shared MCP Session Allows Cross-User API Identity Confusion
Summary
FlowIntel's chatbot component used a single shared MCP (Model Context Protocol) session and process-wide environment variable to pass per-user API credentials to a child process. Because the session and the ambient FLOWINTEL_API_KEY environment variable were shared across all concurrent requests, a user's authenticated identity could be applied to another user's tool invocation, or a stale/ambient credential could be inherited by a new request. Preconditions: - An authenticated user with access to the chatbot interface. - Concurrent or overlapping chatbot tool invocations from different users, or a request made after another user's credentials were written to the process environment. Impact: - A lower-privilege user could read data belonging to a higher-privilege user (e.g., private investigation cases). - A user whose session was invalidated (deleted, password changed, key revoked) could still operate under a previously cached ambient credential. - The parent process environment was mutated with user-specific secrets, creating a persistent credential leak surface. Affected versions: FlowIntel prior to 4.1.0.
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CNA (v2.0.3)
Decision recorded 2026-11-20 14:41 UTC
CWE
  • CWE-290 - Authentication Bypass by Spoofing
  • CWE-200 - Exposure of Sensitive Information to an Unauthorized Actor
Assigner
CIRCL GNA GNA-1
GNA scorecard A 92/100 over 319 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 14:41
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/9eec395cc73c6706362b2c22f1e264406eb8caf8 b2eb7327abee…
Confidence
medium
Commit Subject Patch SHA-256
9eec395cc73c fix: [security] Isolate chatbot MCP API identities b2eb7327abee…
Fix summary

The shared persistent MCP session and process-wide API key mutation were removed. Each chatbot tool invocation now spawns its own short-lived MCP child process with credentials resolved fresh from the authenticated request context. If the user's identity or API key is unavailable, the invocation is rejected before any child process is launched, preventing fallback to ambient or stale credentials. Each transport is deterministically closed on success, failure, or timeout, eliminating residual credential exposure.

Patch summary

Removed the global _mcp_session, _mcp_loop, _mcp_thread, _mcp_current_api_key state and the _ensure_mcp() / _start_mcp_background() / _run_mcp_session() functions that maintained a single shared asyncio event loop and thread. Replaced with _mcp_environment() which resolves the user's API key from the authenticated session on every call and raises RuntimeError if unavailable, and _call_mcp_tool() which opens a per-invocation stdio transport, executes the tool, and closes the transport in all exit paths. Eliminated all os.environ mutations of FLOWINTEL_API_KEY. Added 33 regression tests in tests/chatbot/test_api_identity.py covering concurrent identity isolation, ambient-credential rejection, detached-user handling, key rotation, transport cleanup on failure/timeout, and real stdio child-process lifecycle.

CVSS rationale

AV:N: the chatbot is a network-accessible Flask endpoint. AC:H: exploitation requires concurrent or carefully timed requests from two different users to trigger the shared-session race, or a specific sequence of credential writes to the process environment. AT:N: no manipulation of a separate target is needed. PR:L: the attacker must be an authenticated user with chatbot access. UI:N: no victim interaction beyond their own request is required. VC:H: a user can read another user's sensitive data (private cases, investigations). VI:N, VA:N: no integrity or availability impact is evidenced. SC:N, SI:N, SA:N: no impact on separate components is demonstrated.

Weakness rationale
  • CWE-290 The shared MCP session and process-wide environment variable allowed one user's authenticated identity (API key) to be applied to another user's tool invocation. The per-request identity was not isolated, effectively allowing identity confusion or spoofing between concurrent sessions.
  • CWE-200 Because credentials were shared across requests and written to the process environment, a user could access data (e.g., private cases) belonging to another user whose key was cached in the shared session or ambient environment.
Attack pattern rationale
  • CAPEC-126 The closest available CAPEC. The shared MCP session acted as a single long-lived session whose credentials could be 'hijacked' by a concurrent request from a different user, causing the wrong identity to be used. No CAPEC specifically covers 'shared per-tenant credential state in a multi-user service,' so CAPEC-126 is the best approximation. Uncertainty: the attack does not involve a traditional network-level session token theft but rather a server-side shared-state race.
Assumptions to verify
  • AC:H is assumed because reliable exploitation requires concurrent requests from two different authenticated users or a specific timing window around the shared session restart; a single sequential request may not trigger the identity confusion.
  • PR:L assumes any authenticated user with chatbot access can trigger the issue; the patch does not evidence a higher privilege requirement.
  • VC:H assumes the data accessed (private cases, investigations) constitutes high confidentiality impact; the patch tests reference 'PRIVATE ADMIN INVESTIGATION' as the sensitive resource.
  • CAPEC-126 (Session Hijacking) is the closest available mapping but does not precisely describe a server-side shared-credential race; no CAPEC specifically covers multi-tenant credential isolation failures in a shared in-process session.
  • The affected version range is taken from the explicit analyst context (affected_version: 0, fixed_version: 4.1.0); the patch itself does not encode version boundaries.
  • No evidence of active exploitation, public PoC, or in-the-wild abuse is present in the patch or metadata.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 6 11 medium 6
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-17 10:37 UTC Fix developed Corrective change authored (9eec395cc73c6706362b2c22f1e264406eb8caf8): fix: [security] Isolate chatbot MCP API identities https://github.com/flowintel/flowintel/commit/9eec395cc73c6706362b2c22f1e264406eb8caf8
  2. 2026-11-20 14:41 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/9eec395cc73c6706362b2c22f1e264406eb8caf8

{
  "containers": {
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/chatbot/chatbot_core.py"
          ],
          "product": "flowintel",
          "programFiles": [
            "app/chatbot/chatbot_core.py"
          ],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "FlowIntel\u0027s chatbot component used a single shared MCP (Model Context Protocol) session and process-wide environment variable to pass per-user API credentials to a child process. Because the session and the ambient FLOWINTEL_API_KEY environment variable were shared across all concurrent requests, a user\u0027s authenticated identity could be applied to another user\u0027s tool invocation, or a stale/ambient credential could be inherited by a new request.\n\nPreconditions:\n- An authenticated user with access to the chatbot interface.\n- Concurrent or overlapping chatbot tool invocations from different users, or a request made after another user\u0027s credentials were written to the process environment.\n\nImpact:\n- A lower-privilege user could read data belonging to a higher-privilege user (e.g., private investigation cases).\n- A user whose session was invalidated (deleted, password changed, key revoked) could still operate under a previously cached ambient credential.\n- The parent process environment was mutated with user-specific secrets, creating a persistent credential leak surface.\n\nAffected versions: FlowIntel prior to 4.1.0."
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-126",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-126 Session Hijacking"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "HIGH",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 6,
            "baseSeverity": "MEDIUM",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:H/AT:N/PR:L/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "HIGH",
            "vulnIntegrityImpact": "NONE",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T14:41:59Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-290",
              "description": "CWE-290 Authentication Bypass by Spoofing",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-200",
              "description": "CWE-200 Exposure of Sensitive Information to an Unauthorized Actor",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "orgId": "00000000-0000-4000-8000-000000000000"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/9eec395cc73c6706362b2c22f1e264406eb8caf8"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The shared persistent MCP session and process-wide API key mutation were removed. Each chatbot tool invocation now spawns its own short-lived MCP child process with credentials resolved fresh from the authenticated request context. If the user\u0027s identity or API key is unavailable, the invocation is rejected before any child process is launched, preventing fallback to ambient or stale credentials. Each transport is deterministically closed on success, failure, or timeout, eliminating residual credential exposure."
        }
      ],
      "title": "FlowIntel Chatbot Shared MCP Session Allows Cross-User API Identity Confusion",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "AC:H is assumed because reliable exploitation requires concurrent requests from two different authenticated users or a specific timing window around the shared session restart; a single sequential request may not trigger the identity confusion.",
                  "PR:L assumes any authenticated user with chatbot access can trigger the issue; the patch does not evidence a higher privilege requirement.",
                  "VC:H assumes the data accessed (private cases, investigations) constitutes high confidentiality impact; the patch tests reference \u0027PRIVATE ADMIN INVESTIGATION\u0027 as the sensitive resource.",
                  "CAPEC-126 (Session Hijacking) is the closest available mapping but does not precisely describe a server-side shared-credential race; no CAPEC specifically covers multi-tenant credential isolation failures in a shared in-process session.",
                  "The affected version range is taken from the explicit analyst context (affected_version: 0, fixed_version: 4.1.0); the patch itself does not encode version boundaries.",
                  "No evidence of active exploitation, public PoC, or in-the-wild abuse is present in the patch or metadata."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-126",
                    "rationale": "The closest available CAPEC. The shared MCP session acted as a single long-lived session whose credentials could be \u0027hijacked\u0027 by a concurrent request from a different user, causing the wrong identity to be used. No CAPEC specifically covers \u0027shared per-tenant credential state in a multi-user service,\u0027 so CAPEC-126 is the best approximation. Uncertainty: the attack does not involve a traditional network-level session token theft but rather a server-side shared-state race."
                  }
                ],
                "commit": "9eec395cc73c6706362b2c22f1e264406eb8caf8",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N: the chatbot is a network-accessible Flask endpoint. AC:H: exploitation requires concurrent or carefully timed requests from two different users to trigger the shared-session race, or a specific sequence of credential writes to the process environment. AT:N: no manipulation of a separate target is needed. PR:L: the attacker must be an authenticated user with chatbot access. UI:N: no victim interaction beyond their own request is required. VC:H: a user can read another user\u0027s sensitive data (private cases, investigations). VI:N, VA:N: no integrity or availability impact is evidenced. SC:N, SI:N, SA:N: no impact on separate components is demonstrated.",
                "fixSummary": "The shared persistent MCP session and process-wide API key mutation were removed. Each chatbot tool invocation now spawns its own short-lived MCP child process with credentials resolved fresh from the authenticated request context. If the user\u0027s identity or API key is unavailable, the invocation is rejected before any child process is launched, preventing fallback to ambient or stale credentials. Each transport is deterministically closed on success, failure, or timeout, eliminating residual credential exposure.",
                "generatedAt": "2026-11-20T14:41:59.093131Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 6,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 6
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "b2eb7327abee28d605ef02f5af16f090ce7a643bd63bf2a5836e30539adfbb81",
                "patchSummary": "Removed the global _mcp_session, _mcp_loop, _mcp_thread, _mcp_current_api_key state and the _ensure_mcp() / _start_mcp_background() / _run_mcp_session() functions that maintained a single shared asyncio event loop and thread. Replaced with _mcp_environment() which resolves the user\u0027s API key from the authenticated session on every call and raises RuntimeError if unavailable, and _call_mcp_tool() which opens a per-invocation stdio transport, executes the tool, and closes the transport in all exit paths. Eliminated all os.environ mutations of FLOWINTEL_API_KEY. Added 33 regression tests in tests/chatbot/test_api_identity.py covering concurrent identity isolation, ambient-credential rejection, detached-user handling, key rotation, transport cleanup on failure/timeout, and real stdio child-process lifecycle.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "9eec395cc73c6706362b2c22f1e264406eb8caf8",
                    "date": "Tue, 17 Nov 2026 11:37:34 +0100",
                    "patchSha256": "b2eb7327abee28d605ef02f5af16f090ce7a643bd63bf2a5836e30539adfbb81",
                    "retrievedAt": "2026-11-20T14:41:10.165065Z",
                    "source": "https://github.com/flowintel/flowintel/commit/9eec395cc73c6706362b2c22f1e264406eb8caf8",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/9eec395cc73c6706362b2c22f1e264406eb8caf8",
                    "subject": "fix: [security] Isolate chatbot MCP API identities"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/9eec395cc73c6706362b2c22f1e264406eb8caf8",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T14:41:59Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Isolate chatbot MCP API identities",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-290",
                    "rationale": "The shared MCP session and process-wide environment variable allowed one user\u0027s authenticated identity (API key) to be applied to another user\u0027s tool invocation. The per-request identity was not isolated, effectively allowing identity confusion or spoofing between concurrent sessions."
                  },
                  {
                    "cweId": "CWE-200",
                    "rationale": "Because credentials were shared across requests and written to the process environment, a user could access data (e.g., private cases) belonging to another user whose key was cached in the shared session or ambient environment."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (9eec395cc73c6706362b2c22f1e264406eb8caf8): fix: [security] Isolate chatbot MCP API identities",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/9eec395cc73c6706362b2c22f1e264406eb8caf8"
                    ],
                    "timestamp": "2026-11-17T10:37:34Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/9eec395cc73c6706362b2c22f1e264406eb8caf8"
                    ],
                    "timestamp": "2026-11-20T14:41:10.165065Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20421"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "00000000-0000-4000-8000-000000000000",
    "cveId": "CVE-2026-133853",
    "datePublished": "2026-11-20T14:44:12.578227Z",
    "dateReserved": "2026-11-20T14:44:25.304Z",
    "dateUpdated": "2026-11-20T14:44:26.385701Z",
    "state": "PUBLISHED",
    "vulnId": "GCVE-1-2026-20421"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

GCVE-1-2026-20420 (CVE-2026-133852)

Vulnerability from gna-1 – Published: 2026-11-20 14:36 – Updated: 2026-11-20 14:40
VLAI
Title
flowintel Privileged Case Task Approval Bypass via Missing User Context and Template Creation
Summary
flowintel is a case management application that enforces an approval workflow for tasks in privileged cases. Tasks in such cases are required to start in a Requested status, meaning they must receive explicit approval before becoming active. A logic flaw in the task-creation path allowed tasks to be initialized with a Created status (bypassing the approval step) in two scenarios: - When a task was created from a template and no acting user was supplied, the privileged-case check was skipped entirely, leaving the task in Created status. - When a case was created from a template, all copied tasks were unconditionally assigned the Created status regardless of the case's privileged designation. An authenticated user with the ability to create tasks from templates or to instantiate cases from templates could therefore place tasks directly into an active state within privileged cases, circumventing the intended approval gate. This may expose restricted case content or permit actions that should require supervisory sign-off. Affected versions: flowintel prior to 4.1.0.
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CNA (v2.0.3)
Decision recorded 2026-11-20 14:21 UTC
CWE
Assigner
CIRCL GNA GNA-1
GNA scorecard A 92/100 over 319 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 14:21
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/fd23a3cb577b1ac64b95f076578994188332b2d1 ef4bf88c9ae2…
Confidence
medium
Commit Subject Patch SHA-256
fd23a3cb577b fix: [security] Require initial approval for privileged ef4bf88c9ae2…
Fix summary

The approval-workflow enforcement for privileged cases is now applied unconditionally at task creation time. The privileged-case check no longer depends on the presence of an acting user, and template-based case creation explicitly assigns the configured Requested status to tasks when the target case is privileged. Ordinary (non-privileged) cases retain the default Created status, preserving existing behavior.

Patch summary

Two one-line changes were made. In app/case/common_core.py, the condition guarding the Requested-status assignment was changed from 'if current_user and case.privileged_case' to 'if case.privileged_case', removing the dependency on a non-null user. In app/templating/TemplateCase.py, the hardcoded 'status_id=1' in the task-copy constructor was replaced with a conditional expression that selects the configured TASK_REQUESTED status when the case is privileged and falls back to 1 otherwise.

CVSS rationale

AV:N: flowintel is a web application accessible over the network. AC:L: the bypass is triggered simply by the absence of a user context or by using the template-creation path; no race condition or special timing is needed. AT:N: no manipulation of a separate attack target is required. PR:L: the attacker must be an authenticated user with permission to create tasks from templates or instantiate cases from templates. UI:N: no victim interaction is required. VC:L: tasks in privileged cases become immediately accessible without approval, potentially exposing restricted case content. VI:L: the integrity of the approval workflow is compromised; tasks are in an unauthorized state. VA:N: no availability impact. SC/SI/SA:N: no secondary impacts on other systems are evidenced.

Weakness rationale
  • CWE-862 The approval workflow (a form of authorization gate) for privileged-case tasks was not enforced under certain code paths (no user context, template-based creation), allowing tasks to skip the required approval step. The narrowest defensible CWE is Missing Authorization because the security control (approval) was simply absent in those paths.
Attack pattern rationale
  • CAPEC-108 The vulnerability is triggered by the absence of a user parameter, which causes the privileged-case authorization check to be skipped. An attacker who can influence whether a user context is supplied (e.g., via an API endpoint that does not enforce authentication for this code path) effectively tampers with the parameter to bypass the control. This is the closest CAPEC available; the mapping is uncertain because the primary defect is a missing check rather than active parameter manipulation, but CAPEC-108 best captures the 'missing/absent parameter leads to skipped validation' pattern.
Assumptions to verify
  • The 'privileged case' designation implies a higher-sensitivity classification requiring supervisory approval; the exact business meaning is inferred from the code context and commit message.
  • PR:L assumes the attacker needs an authenticated account with template-creation permissions; the patch does not specify the exact role or permission required.
  • CAPEC-108 is the closest available mapping; the primary defect is a missing conditional check rather than active parameter manipulation, so the mapping is approximate.
  • The affected version range '0 to <4.1.0' is taken from the explicit analyst context; the patch itself does not encode version boundaries.
  • VC:L assumes that bypassing the approval step exposes task content that would otherwise be restricted; the patch does not detail what content is protected.
  • The commit message references 'Tool: Codex' which is noted as the AI tool used for implementation, not as a credit.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 6 11 medium 6
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-17 10:02 UTC Fix developed Corrective change authored (fd23a3cb577b1ac64b95f076578994188332b2d1): fix: [security] Require initial approval for privileged https://github.com/flowintel/flowintel/commit/fd23a3cb577b1ac64b95f076578994188332b2d1
  2. 2026-11-20 14:21 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/fd23a3cb577b1ac64b95f076578994188332b2d1

{
  "containers": {
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/case/common_core.py",
            "app/templating/TemplateCase.py"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "flowintel is a case management application that enforces an approval workflow for tasks in privileged cases. Tasks in such cases are required to start in a Requested status, meaning they must receive explicit approval before becoming active.\n\nA logic flaw in the task-creation path allowed tasks to be initialized with a Created status (bypassing the approval step) in two scenarios:\n\n- When a task was created from a template and no acting user was supplied, the privileged-case check was skipped entirely, leaving the task in Created status.\n- When a case was created from a template, all copied tasks were unconditionally assigned the Created status regardless of the case\u0027s privileged designation.\n\nAn authenticated user with the ability to create tasks from templates or to instantiate cases from templates could therefore place tasks directly into an active state within privileged cases, circumventing the intended approval gate. This may expose restricted case content or permit actions that should require supervisory sign-off.\n\nAffected versions: flowintel prior to 4.1.0."
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-108",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-108 Parameter Tampering"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 5.3,
            "baseSeverity": "MEDIUM",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:L/VA:N/SC:N/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "LOW",
            "vulnIntegrityImpact": "LOW",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T14:21:30Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-862",
              "description": "CWE-862 Missing Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "orgId": "00000000-0000-4000-8000-000000000000"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/fd23a3cb577b1ac64b95f076578994188332b2d1"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The approval-workflow enforcement for privileged cases is now applied unconditionally at task creation time. The privileged-case check no longer depends on the presence of an acting user, and template-based case creation explicitly assigns the configured Requested status to tasks when the target case is privileged. Ordinary (non-privileged) cases retain the default Created status, preserving existing behavior."
        }
      ],
      "title": "flowintel Privileged Case Task Approval Bypass via Missing User Context and Template Creation",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The \u0027privileged case\u0027 designation implies a higher-sensitivity classification requiring supervisory approval; the exact business meaning is inferred from the code context and commit message.",
                  "PR:L assumes the attacker needs an authenticated account with template-creation permissions; the patch does not specify the exact role or permission required.",
                  "CAPEC-108 is the closest available mapping; the primary defect is a missing conditional check rather than active parameter manipulation, so the mapping is approximate.",
                  "The affected version range \u00270 to \u003c4.1.0\u0027 is taken from the explicit analyst context; the patch itself does not encode version boundaries.",
                  "VC:L assumes that bypassing the approval step exposes task content that would otherwise be restricted; the patch does not detail what content is protected.",
                  "The commit message references \u0027Tool: Codex\u0027 which is noted as the AI tool used for implementation, not as a credit."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-108",
                    "rationale": "The vulnerability is triggered by the absence of a user parameter, which causes the privileged-case authorization check to be skipped. An attacker who can influence whether a user context is supplied (e.g., via an API endpoint that does not enforce authentication for this code path) effectively tampers with the parameter to bypass the control. This is the closest CAPEC available; the mapping is uncertain because the primary defect is a missing check rather than active parameter manipulation, but CAPEC-108 best captures the \u0027missing/absent parameter leads to skipped validation\u0027 pattern."
                  }
                ],
                "commit": "fd23a3cb577b1ac64b95f076578994188332b2d1",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N: flowintel is a web application accessible over the network. AC:L: the bypass is triggered simply by the absence of a user context or by using the template-creation path; no race condition or special timing is needed. AT:N: no manipulation of a separate attack target is required. PR:L: the attacker must be an authenticated user with permission to create tasks from templates or instantiate cases from templates. UI:N: no victim interaction is required. VC:L: tasks in privileged cases become immediately accessible without approval, potentially exposing restricted case content. VI:L: the integrity of the approval workflow is compromised; tasks are in an unauthorized state. VA:N: no availability impact. SC/SI/SA:N: no secondary impacts on other systems are evidenced.",
                "fixSummary": "The approval-workflow enforcement for privileged cases is now applied unconditionally at task creation time. The privileged-case check no longer depends on the presence of an acting user, and template-based case creation explicitly assigns the configured Requested status to tasks when the target case is privileged. Ordinary (non-privileged) cases retain the default Created status, preserving existing behavior.",
                "generatedAt": "2026-11-20T14:21:30.590126Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 6,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 6
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "ef4bf88c9ae2f0d59b98cffbeb804fbbbeeb8017452a730e8c60200aa01ca687",
                "patchSummary": "Two one-line changes were made. In app/case/common_core.py, the condition guarding the Requested-status assignment was changed from \u0027if current_user and case.privileged_case\u0027 to \u0027if case.privileged_case\u0027, removing the dependency on a non-null user. In app/templating/TemplateCase.py, the hardcoded \u0027status_id=1\u0027 in the task-copy constructor was replaced with a conditional expression that selects the configured TASK_REQUESTED status when the case is privileged and falls back to 1 otherwise.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "fd23a3cb577b1ac64b95f076578994188332b2d1",
                    "date": "Tue, 17 Nov 2026 11:02:12 +0100",
                    "patchSha256": "ef4bf88c9ae2f0d59b98cffbeb804fbbbeeb8017452a730e8c60200aa01ca687",
                    "retrievedAt": "2026-11-20T14:21:01.216442Z",
                    "source": "https://github.com/flowintel/flowintel/commit/fd23a3cb577b1ac64b95f076578994188332b2d1",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/fd23a3cb577b1ac64b95f076578994188332b2d1",
                    "subject": "fix: [security] Require initial approval for privileged"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/fd23a3cb577b1ac64b95f076578994188332b2d1",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T14:21:30Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Require initial approval for privileged",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-862",
                    "rationale": "The approval workflow (a form of authorization gate) for privileged-case tasks was not enforced under certain code paths (no user context, template-based creation), allowing tasks to skip the required approval step. The narrowest defensible CWE is Missing Authorization because the security control (approval) was simply absent in those paths."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (fd23a3cb577b1ac64b95f076578994188332b2d1): fix: [security] Require initial approval for privileged",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/fd23a3cb577b1ac64b95f076578994188332b2d1"
                    ],
                    "timestamp": "2026-11-17T10:02:12Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/fd23a3cb577b1ac64b95f076578994188332b2d1"
                    ],
                    "timestamp": "2026-11-20T14:21:01.216442Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20420"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "00000000-0000-4000-8000-000000000000",
    "cveId": "CVE-2026-133852",
    "datePublished": "2026-11-20T14:36:10.514611Z",
    "dateReserved": "2026-11-20T14:40:19.091Z",
    "dateUpdated": "2026-11-20T14:40:19.190484Z",
    "state": "PUBLISHED",
    "vulnId": "GCVE-1-2026-20420"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

GCVE-1-2026-20419 (CVE-2026-133851)

Vulnerability from gna-1 – Published: 2026-11-20 14:36 – Updated: 2026-11-20 14:40
VLAI
Title
FlowIntel Alerting: Missing Private Case Authorization Check Allows Connector Owners to Access Revoked Case Data
Summary
FlowIntel's alerting module contained an authorization bypass in the alert access control logic. When an alert was linked to a private case, the system relied solely on connector ownership (user or organization) to determine access, without verifying that the requesting user still held valid membership in the associated private case. This meant that a user who was a connector owner of an alert could continue to read alert details, review history, raw payloads, and action comments, and could perform review-status updates or case-creation actions, even after their access to the linked private case had been revoked or they were never a member of that case. Additionally, a time-of-check to time-of-use (TOCTOU) gap existed: after a review-status update or a competing case-creation claim was committed, the response could disclose private case identifiers or review history to a user whose access had been concurrently removed by another session linking the alert to a private case. Impact: - Unauthorized disclosure of private case data (review comments, raw payloads, action history) through alert endpoints. - Unauthorized modification of alert review status and creation of cases from alerts the user should no longer access. - Leakage of a competing case identifier to an unauthorized user during concurrent case-creation races. Affected versions: FlowIntel prior to 4.1.0.
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CNA (v2.0.3)
Decision recorded 2026-11-20 14:20 UTC
CWE
  • CWE-862 - Missing Authorization
  • CWE-362 - Concurrent Execution using Shared Resource with Inadequate Locking
Assigner
CIRCL GNA GNA-1
GNA scorecard A 92/100 over 319 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 14:20
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/68d85c78e464f023f9f77af9af39634d3b811fff e8226318f1d6…
Confidence
medium
Commit Subject Patch SHA-256
68d85c78e464 fix: [security] Enforce private case access for alert reviews e8226318f1d6…
Fix summary

The fix enforces private case visibility as a mandatory gate before any connector-ownership check can grant alert access. If an alert is linked to a private case and the user cannot view that case, access is denied regardless of ownership. A post-commit re-verification is added after review-status updates and after detecting a competing case-creation claim, ensuring that concurrent association changes do not result in unauthorized data disclosure in the response.

Patch summary

In can_user_view_alert, the private-case membership check was moved before the connector-ownership checks and made unconditional (no longer gated on case.is_private), so any user lacking case access is denied even if they own the connector. In update_review_status, a can_user_view_alert re-check is performed after the database commit and before returning the serialized alert, returning 403 if access was concurrently lost. In create_case_from_alert, after a competing claim is detected, a can_user_view_alert re-check gates the response that would otherwise return the winner's case_id. The unused import get_present_in_case was removed. 33 regression tests were added covering revoked access, connector scope, concurrent case-link transitions, and authorized-user preservation.

CVSS rationale

AV:N: The vulnerability is exploitable over the network via the FlowIntel web API. AC:L: No special race condition or complex setup is required; a connector owner whose case access was revoked can simply call the alert endpoints. AT:N: No manipulation of the target is needed. PR:L: The attacker must be an authenticated user with connector ownership of the alert. UI:N: No victim interaction is required. VC:L: Confidential impact is limited to alert-level data (review comments, raw payload, action history) rather than full case contents. VI:L: The attacker can modify review status and create cases, but cannot alter case data directly. VA:N: No availability impact. SC/SI/SA:N: No impact on other systems.

Weakness rationale
  • CWE-862 The alert access function did not verify private case membership before granting access based on connector ownership. The authorization check for case-level access was either absent or insufficiently ordered, allowing a user with connector ownership to bypass the private case boundary.
  • CWE-362 The review-status update and case-creation paths did not re-verify authorization after the database commit, creating a TOCTOU window where a concurrent session could link the alert to a private case and the response would still disclose case data to an unauthorized user.
Attack pattern rationale
  • CAPEC-114 The attacker directly requests alert API endpoints (review, case-creation, read) using a known alert identifier. Because the authorization logic did not enforce private case membership, the server returned sensitive case data and accepted state-changing actions. This is the closest CAPEC to the observed pattern of accessing a resource by directly specifying its identifier without the required authorization context. Uncertainty: the attack does not involve guessing URLs but rather exploiting a logic flaw in an existing authorized endpoint; however, CAPEC-114 remains the best available match for direct resource access bypassing intended access controls.
Assumptions to verify
  • The affected version range is inferred from the explicit context (affected_version '0', fixed_version '4.1.0'); no explicit changelog or release notes are available to confirm the exact first vulnerable release.
  • The CAPEC-114 mapping is the closest available pattern; the actual attack is a logic flaw in authorization ordering rather than classic forced browsing, but no more specific CAPEC exists for missing authorization checks in API endpoints.
  • CVSS PR:L assumes the attacker must be an authenticated user with connector ownership; if the product allows unauthenticated access to alert endpoints, PR would be N, but the patch evidence (session-based auth, API keys in tests) supports authenticated access.
  • The TOCTOU component (CWE-362) requires a concurrent session to link the alert to a private case at the precise moment of the response; this is a secondary vector and the primary issue is the missing authorization check (CWE-862).
  • The commit message references 'Tool: Codex' which is noted as the AI tool used to assist in development; no credit is assigned for this as it is not a person or organization.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 7 11 medium 5
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-17 09:50 UTC Fix developed Corrective change authored (68d85c78e464f023f9f77af9af39634d3b811fff): fix: [security] Enforce private case access for alert reviews https://github.com/flowintel/flowintel/commit/68d85c78e464f023f9f77af9af39634d3b811fff
  2. 2026-11-20 14:20 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/68d85c78e464f023f9f77af9af39634d3b811fff

{
  "containers": {
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/alerting/alerting_core.py"
          ],
          "product": "flowintel",
          "programFiles": [
            "app/alerting/alerting_core.py"
          ],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "FlowIntel\u0027s alerting module contained an authorization bypass in the alert access control logic. When an alert was linked to a private case, the system relied solely on connector ownership (user or organization) to determine access, without verifying that the requesting user still held valid membership in the associated private case.\n\nThis meant that a user who was a connector owner of an alert could continue to read alert details, review history, raw payloads, and action comments, and could perform review-status updates or case-creation actions, even after their access to the linked private case had been revoked or they were never a member of that case.\n\nAdditionally, a time-of-check to time-of-use (TOCTOU) gap existed: after a review-status update or a competing case-creation claim was committed, the response could disclose private case identifiers or review history to a user whose access had been concurrently removed by another session linking the alert to a private case.\n\nImpact:\n- Unauthorized disclosure of private case data (review comments, raw payloads, action history) through alert endpoints.\n- Unauthorized modification of alert review status and creation of cases from alerts the user should no longer access.\n- Leakage of a competing case identifier to an unauthorized user during concurrent case-creation races.\n\nAffected versions: FlowIntel prior to 4.1.0."
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-114",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-114 Forced Browsing"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 5.3,
            "baseSeverity": "MEDIUM",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:L/VA:N/SC:N/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "LOW",
            "vulnIntegrityImpact": "LOW",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T14:20:59Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-862",
              "description": "CWE-862 Missing Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-362",
              "description": "CWE-362 Concurrent Execution using Shared Resource with Inadequate Locking",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "orgId": "00000000-0000-4000-8000-000000000000"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/68d85c78e464f023f9f77af9af39634d3b811fff"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The fix enforces private case visibility as a mandatory gate before any connector-ownership check can grant alert access. If an alert is linked to a private case and the user cannot view that case, access is denied regardless of ownership. A post-commit re-verification is added after review-status updates and after detecting a competing case-creation claim, ensuring that concurrent association changes do not result in unauthorized data disclosure in the response."
        }
      ],
      "title": "FlowIntel Alerting: Missing Private Case Authorization Check Allows Connector Owners to Access Revoked Case Data",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected version range is inferred from the explicit context (affected_version \u00270\u0027, fixed_version \u00274.1.0\u0027); no explicit changelog or release notes are available to confirm the exact first vulnerable release.",
                  "The CAPEC-114 mapping is the closest available pattern; the actual attack is a logic flaw in authorization ordering rather than classic forced browsing, but no more specific CAPEC exists for missing authorization checks in API endpoints.",
                  "CVSS PR:L assumes the attacker must be an authenticated user with connector ownership; if the product allows unauthenticated access to alert endpoints, PR would be N, but the patch evidence (session-based auth, API keys in tests) supports authenticated access.",
                  "The TOCTOU component (CWE-362) requires a concurrent session to link the alert to a private case at the precise moment of the response; this is a secondary vector and the primary issue is the missing authorization check (CWE-862).",
                  "The commit message references \u0027Tool: Codex\u0027 which is noted as the AI tool used to assist in development; no credit is assigned for this as it is not a person or organization."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-114",
                    "rationale": "The attacker directly requests alert API endpoints (review, case-creation, read) using a known alert identifier. Because the authorization logic did not enforce private case membership, the server returned sensitive case data and accepted state-changing actions. This is the closest CAPEC to the observed pattern of accessing a resource by directly specifying its identifier without the required authorization context. Uncertainty: the attack does not involve guessing URLs but rather exploiting a logic flaw in an existing authorized endpoint; however, CAPEC-114 remains the best available match for direct resource access bypassing intended access controls."
                  }
                ],
                "commit": "68d85c78e464f023f9f77af9af39634d3b811fff",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N: The vulnerability is exploitable over the network via the FlowIntel web API. AC:L: No special race condition or complex setup is required; a connector owner whose case access was revoked can simply call the alert endpoints. AT:N: No manipulation of the target is needed. PR:L: The attacker must be an authenticated user with connector ownership of the alert. UI:N: No victim interaction is required. VC:L: Confidential impact is limited to alert-level data (review comments, raw payload, action history) rather than full case contents. VI:L: The attacker can modify review status and create cases, but cannot alter case data directly. VA:N: No availability impact. SC/SI/SA:N: No impact on other systems.",
                "fixSummary": "The fix enforces private case visibility as a mandatory gate before any connector-ownership check can grant alert access. If an alert is linked to a private case and the user cannot view that case, access is denied regardless of ownership. A post-commit re-verification is added after review-status updates and after detecting a competing case-creation claim, ensuring that concurrent association changes do not result in unauthorized data disclosure in the response.",
                "generatedAt": "2026-11-20T14:20:59.693658Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 5,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 7
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "e8226318f1d6462435435a1293e4a5091a59c7cd5671dbd98013ff3ae14d856a",
                "patchSummary": "In can_user_view_alert, the private-case membership check was moved before the connector-ownership checks and made unconditional (no longer gated on case.is_private), so any user lacking case access is denied even if they own the connector. In update_review_status, a can_user_view_alert re-check is performed after the database commit and before returning the serialized alert, returning 403 if access was concurrently lost. In create_case_from_alert, after a competing claim is detected, a can_user_view_alert re-check gates the response that would otherwise return the winner\u0027s case_id. The unused import get_present_in_case was removed. 33 regression tests were added covering revoked access, connector scope, concurrent case-link transitions, and authorized-user preservation.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "68d85c78e464f023f9f77af9af39634d3b811fff",
                    "date": "Tue, 17 Nov 2026 10:50:38 +0100",
                    "patchSha256": "e8226318f1d6462435435a1293e4a5091a59c7cd5671dbd98013ff3ae14d856a",
                    "retrievedAt": "2026-11-20T14:20:26.159139Z",
                    "source": "https://github.com/flowintel/flowintel/commit/68d85c78e464f023f9f77af9af39634d3b811fff",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/68d85c78e464f023f9f77af9af39634d3b811fff",
                    "subject": "fix: [security] Enforce private case access for alert reviews"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/68d85c78e464f023f9f77af9af39634d3b811fff",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T14:20:59Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Enforce private case access for alert reviews",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-862",
                    "rationale": "The alert access function did not verify private case membership before granting access based on connector ownership. The authorization check for case-level access was either absent or insufficiently ordered, allowing a user with connector ownership to bypass the private case boundary."
                  },
                  {
                    "cweId": "CWE-362",
                    "rationale": "The review-status update and case-creation paths did not re-verify authorization after the database commit, creating a TOCTOU window where a concurrent session could link the alert to a private case and the response would still disclose case data to an unauthorized user."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (68d85c78e464f023f9f77af9af39634d3b811fff): fix: [security] Enforce private case access for alert reviews",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/68d85c78e464f023f9f77af9af39634d3b811fff"
                    ],
                    "timestamp": "2026-11-17T09:50:38Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/68d85c78e464f023f9f77af9af39634d3b811fff"
                    ],
                    "timestamp": "2026-11-20T14:20:26.159139Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20419"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "00000000-0000-4000-8000-000000000000",
    "cveId": "CVE-2026-133851",
    "datePublished": "2026-11-20T14:36:06.387623Z",
    "dateReserved": "2026-11-20T14:40:00.274Z",
    "dateUpdated": "2026-11-20T14:40:00.368848Z",
    "state": "PUBLISHED",
    "vulnId": "GCVE-1-2026-20419"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

GCVE-1-2026-20418 (CVE-2026-133850)

Vulnerability from gna-1 – Published: 2026-11-20 14:36 – Updated: 2026-11-20 14:39
VLAI
Title
flowintel: Stale Authorization Allows Notification Disclosure After Case Access Revocation
Summary
flowintel contains an authorization flaw in its recurring reminder notification subsystem. When a user is subscribed to recurring notifications for a case, the system does not re-verify the recipient's current access rights to that case before generating or displaying the notification. As a result, a user who previously had access to a private case and later had that access revoked can still: - Receive recurring reminder notifications containing case-related content. - View stored reminder notifications in their notification list, including notification counts, even though they no longer have permission to view the underlying case. The vulnerability affects both the notification creation path (recurring reminder generation) and the notification display path (listing and counting stored reminders). The issue arises because the authorization check performed at subscription time is not re-applied at delivery or retrieval time. Security impact is limited to information disclosure: an unauthorized user may observe case titles, reminder messages, and associated metadata that they no longer have permission to access. No integrity or availability impact is indicated by the patch. Affected versions: flowintel prior to 4.1.0.
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CNA (v2.0.3)
Decision recorded 2026-11-20 14:20 UTC
CWE
Assigner
CIRCL GNA GNA-1
GNA scorecard A 92/100 over 319 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 14:20
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/af6e91c8ac63c7ac0860325582bc2561ca5343ea 433c4d2d3f96…
Confidence
medium
Commit Subject Patch SHA-256
af6e91c8ac63 fix: [security] Recheck private case access for recurring 433c4d2d3f96…
Fix summary

The fix introduces an authorization re-check at two points: (1) before creating a recurring reminder notification, the system now verifies the recipient's current case visibility and skips delivery if access has been revoked; (2) when listing or counting stored notifications, the system now filters out case-related reminder notifications for which the viewing user no longer has case access. This ensures that both the creation and display paths enforce current authorization state rather than relying on a one-time check at subscription time.

Patch summary

In recurring_notification.py, the create_notification function now imports can_view_case and, for each recurring notification subscriber, fetches the User record and calls can_view_case(case, recipient); if the check fails, the notification is skipped via continue. In notification_core.py, the DEADLINE_ICONS tuple is extended to include the clock icon (fa-solid fa-clock), a new helper _is_case_deadline_notification is added to identify case-related deadline/reminder notifications by icon or notification_type, and _visible_deadline_notifications is updated to use this helper so that reminders for inaccessible cases are excluded from the returned list. The test file is updated to create a real Case object for the reminder notification fixture instead of using a hardcoded case_id.

CVSS rationale

AV:N: The application is a web-based platform accessed over the network. AC:L: No special race condition or complex setup is required; the user simply needs to have previously had case access and then have it revoked. AT:N: No manipulation of the target is needed. PR:L: An authenticated user with a valid account and prior case subscription is required. UI:N: Notifications are delivered and displayed automatically without requiring victim interaction. VC:L: Limited confidentiality impact; case-related notification content (titles, reminder messages) is exposed to a user who no longer has case access. VI:N, VA:N: No integrity or availability impact is evidenced. SC:N, SI:N, SA:N: The affected data resides within the same application scope; no cross-scope impact is indicated.

Weakness rationale
  • CWE-285 The system performed an authorization check at subscription time but failed to re-enforce it at notification delivery and display time. The authorization decision was not re-evaluated when the user's access to the case changed, resulting in continued access to case-related notification content. This is a failure to properly enforce authorization at the point of access rather than a complete absence of authorization logic.
  • CWE-862 In the recurring notification creation path, no authorization check existed at all before generating the notification for a subscriber. The check was entirely absent, not merely insufficient, making CWE-862 a defensible secondary mapping for the creation path specifically.
Attack pattern rationale
  • CAPEC-66 The notification subsystem operated with a lower security enforcement level than the case-access subsystem: it did not enforce the same access-control checks that governed case visibility. An attacker (or a user whose access was legitimately revoked) could exploit this misconfiguration to continue receiving case-related notification content. This is the closest CAPEC to the pattern of a component not enforcing the security level configured elsewhere in the application. The mapping is uncertain because no CAPEC specifically describes stale-authorization or missing re-authorization at delivery time; CAPEC-66 is selected as the best available match for a security-level enforcement gap between components.
Assumptions to verify
  • The affected version range is inferred from the explicit context (affected_version: 0, fixed_version: 4.1.0); the patch itself does not contain version strings.
  • The CAPEC-66 mapping is the closest available match; no CAPEC specifically covers stale-authorization or missing re-authorization at notification delivery time. This uncertainty is noted in the CAPEC rationale.
  • CVSS PR:L assumes the attacker is an authenticated user who previously had case access; the patch does not specify whether unauthenticated access is possible.
  • CVSS VC:L assumes the disclosed information is limited to notification content (case titles, reminder messages) rather than full case data; the patch only addresses notification visibility.
  • The patch does not indicate whether the vulnerability is exploitable remotely or requires local access; AV:N is assumed based on the web-application nature of the product.
  • The commit message references 'Tool: Codex' which is noted as the AI tool used for implementation but is not credited as a vulnerability finder or reporter.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 6 11 medium 6
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-17 09:16 UTC Fix developed Corrective change authored (af6e91c8ac63c7ac0860325582bc2561ca5343ea): fix: [security] Recheck private case access for recurring https://github.com/flowintel/flowintel/commit/af6e91c8ac63c7ac0860325582bc2561ca5343ea
  2. 2026-11-20 14:19 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/af6e91c8ac63c7ac0860325582bc2561ca5343ea

{
  "containers": {
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/bin/recurring_notification.py",
            "app/notification/notification_core.py"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "flowintel contains an authorization flaw in its recurring reminder notification subsystem. When a user is subscribed to recurring notifications for a case, the system does not re-verify the recipient\u0027s current access rights to that case before generating or displaying the notification.\n\nAs a result, a user who previously had access to a private case and later had that access revoked can still:\n- Receive recurring reminder notifications containing case-related content.\n- View stored reminder notifications in their notification list, including notification counts, even though they no longer have permission to view the underlying case.\n\nThe vulnerability affects both the notification creation path (recurring reminder generation) and the notification display path (listing and counting stored reminders). The issue arises because the authorization check performed at subscription time is not re-applied at delivery or retrieval time.\n\nSecurity impact is limited to information disclosure: an unauthorized user may observe case titles, reminder messages, and associated metadata that they no longer have permission to access. No integrity or availability impact is indicated by the patch.\n\nAffected versions: flowintel prior to 4.1.0."
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-66",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-66 Exploiting Incorrectly Configured Security Levels"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 5.3,
            "baseSeverity": "MEDIUM",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:N/VA:N/SC:N/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "LOW",
            "vulnIntegrityImpact": "NONE",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T14:20:25Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-285",
              "description": "CWE-285 Improper Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-862",
              "description": "CWE-862 Missing Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "orgId": "00000000-0000-4000-8000-000000000000"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/af6e91c8ac63c7ac0860325582bc2561ca5343ea"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The fix introduces an authorization re-check at two points: (1) before creating a recurring reminder notification, the system now verifies the recipient\u0027s current case visibility and skips delivery if access has been revoked; (2) when listing or counting stored notifications, the system now filters out case-related reminder notifications for which the viewing user no longer has case access. This ensures that both the creation and display paths enforce current authorization state rather than relying on a one-time check at subscription time."
        }
      ],
      "title": "flowintel: Stale Authorization Allows Notification Disclosure After Case Access Revocation",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected version range is inferred from the explicit context (affected_version: 0, fixed_version: 4.1.0); the patch itself does not contain version strings.",
                  "The CAPEC-66 mapping is the closest available match; no CAPEC specifically covers stale-authorization or missing re-authorization at notification delivery time. This uncertainty is noted in the CAPEC rationale.",
                  "CVSS PR:L assumes the attacker is an authenticated user who previously had case access; the patch does not specify whether unauthenticated access is possible.",
                  "CVSS VC:L assumes the disclosed information is limited to notification content (case titles, reminder messages) rather than full case data; the patch only addresses notification visibility.",
                  "The patch does not indicate whether the vulnerability is exploitable remotely or requires local access; AV:N is assumed based on the web-application nature of the product.",
                  "The commit message references \u0027Tool: Codex\u0027 which is noted as the AI tool used for implementation but is not credited as a vulnerability finder or reporter."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-66",
                    "rationale": "The notification subsystem operated with a lower security enforcement level than the case-access subsystem: it did not enforce the same access-control checks that governed case visibility. An attacker (or a user whose access was legitimately revoked) could exploit this misconfiguration to continue receiving case-related notification content. This is the closest CAPEC to the pattern of a component not enforcing the security level configured elsewhere in the application. The mapping is uncertain because no CAPEC specifically describes stale-authorization or missing re-authorization at delivery time; CAPEC-66 is selected as the best available match for a security-level enforcement gap between components."
                  }
                ],
                "commit": "af6e91c8ac63c7ac0860325582bc2561ca5343ea",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N: The application is a web-based platform accessed over the network. AC:L: No special race condition or complex setup is required; the user simply needs to have previously had case access and then have it revoked. AT:N: No manipulation of the target is needed. PR:L: An authenticated user with a valid account and prior case subscription is required. UI:N: Notifications are delivered and displayed automatically without requiring victim interaction. VC:L: Limited confidentiality impact; case-related notification content (titles, reminder messages) is exposed to a user who no longer has case access. VI:N, VA:N: No integrity or availability impact is evidenced. SC:N, SI:N, SA:N: The affected data resides within the same application scope; no cross-scope impact is indicated.",
                "fixSummary": "The fix introduces an authorization re-check at two points: (1) before creating a recurring reminder notification, the system now verifies the recipient\u0027s current case visibility and skips delivery if access has been revoked; (2) when listing or counting stored notifications, the system now filters out case-related reminder notifications for which the viewing user no longer has case access. This ensures that both the creation and display paths enforce current authorization state rather than relying on a one-time check at subscription time.",
                "generatedAt": "2026-11-20T14:20:25.637194Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 6,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 6
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "433c4d2d3f9663b3d5b556b1d1236ba7ada4e9ef128b0cca9f53b69d8d1a0e1c",
                "patchSummary": "In recurring_notification.py, the create_notification function now imports can_view_case and, for each recurring notification subscriber, fetches the User record and calls can_view_case(case, recipient); if the check fails, the notification is skipped via continue. In notification_core.py, the DEADLINE_ICONS tuple is extended to include the clock icon (fa-solid fa-clock), a new helper _is_case_deadline_notification is added to identify case-related deadline/reminder notifications by icon or notification_type, and _visible_deadline_notifications is updated to use this helper so that reminders for inaccessible cases are excluded from the returned list. The test file is updated to create a real Case object for the reminder notification fixture instead of using a hardcoded case_id.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "af6e91c8ac63c7ac0860325582bc2561ca5343ea",
                    "date": "Tue, 17 Nov 2026 10:16:17 +0100",
                    "patchSha256": "433c4d2d3f9663b3d5b556b1d1236ba7ada4e9ef128b0cca9f53b69d8d1a0e1c",
                    "retrievedAt": "2026-11-20T14:19:53.580074Z",
                    "source": "https://github.com/flowintel/flowintel/commit/af6e91c8ac63c7ac0860325582bc2561ca5343ea",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/af6e91c8ac63c7ac0860325582bc2561ca5343ea",
                    "subject": "fix: [security] Recheck private case access for recurring"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/af6e91c8ac63c7ac0860325582bc2561ca5343ea",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T14:20:25Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Recheck private case access for recurring",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-285",
                    "rationale": "The system performed an authorization check at subscription time but failed to re-enforce it at notification delivery and display time. The authorization decision was not re-evaluated when the user\u0027s access to the case changed, resulting in continued access to case-related notification content. This is a failure to properly enforce authorization at the point of access rather than a complete absence of authorization logic."
                  },
                  {
                    "cweId": "CWE-862",
                    "rationale": "In the recurring notification creation path, no authorization check existed at all before generating the notification for a subscriber. The check was entirely absent, not merely insufficient, making CWE-862 a defensible secondary mapping for the creation path specifically."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (af6e91c8ac63c7ac0860325582bc2561ca5343ea): fix: [security] Recheck private case access for recurring",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/af6e91c8ac63c7ac0860325582bc2561ca5343ea"
                    ],
                    "timestamp": "2026-11-17T09:16:17Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/af6e91c8ac63c7ac0860325582bc2561ca5343ea"
                    ],
                    "timestamp": "2026-11-20T14:19:53.580074Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20418"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "00000000-0000-4000-8000-000000000000",
    "cveId": "CVE-2026-133850",
    "datePublished": "2026-11-20T14:36:03.890014Z",
    "dateReserved": "2026-11-20T14:39:45.487Z",
    "dateUpdated": "2026-11-20T14:39:45.578853Z",
    "state": "PUBLISHED",
    "vulnId": "GCVE-1-2026-20418"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

GCVE-1-2026-20417 (CVE-2026-133849)

Vulnerability from gna-1 – Published: 2026-11-20 14:35 – Updated: 2026-11-20 14:39
VLAI
Title
flowintel: Insecure Direct Object Reference in Task Calendar Access Control
Summary
A vulnerability exists in the flowintel application's calendar functionality that allows unauthorized access to private case data. The task calendar queries failed to enforce visibility checks on the parent case associated with tasks. Specifically, tasks assigned to a user or belonging to the user's organization were returned in calendar views (including browser and API ICS feeds) without verifying whether the user had permission to view the parent case. This allowed users to potentially discover and access details of private cases they were not authorized to view, leading to information disclosure. The issue affected the default task visibility logic and the assigned-only task queries. The fix ensures that case visibility is rechecked in addition to task assignment or organization membership, preventing the exposure of private case data in calendar outputs. Affected versions: flowintel < 4.1.0
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CNA (v2.0.3)
Decision recorded 2026-11-20 14:19 UTC
CWE
  • CWE-639 - Authorization Bypass Through User-Controlled Key
Assigner
CIRCL GNA GNA-1
GNA scorecard A 92/100 over 319 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 14:19
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/98196d0523d4e1136014574ff0b526da1d92381b 618096dd6233…
Confidence
medium
Commit Subject Patch SHA-256
98196d0523d4 fix: [security] Recheck private case access in task calendars 618096dd6233…
Fix summary

The vulnerability is remediated by adding explicit case visibility checks to the task calendar queries. The fix ensures that tasks are only visible if the user has permission to view the parent case, in addition to being assigned to the task or belonging to the same organization. This prevents the disclosure of private case data in calendar views and ICS feeds.

Patch summary

The patch modifies the `_tasks_visible_to_user_query` and `_tasks_assigned_to_user_query` functions in `app/calendar/calendar_core.py`. It replaces the previous outer join and filter logic with subqueries for task assignment and organization membership. A new filter condition is added to ensure that the case is not private or that the user has membership in the case's organization. This enforces case-level access control in calendar task queries.

CVSS rationale

The vulnerability is exploitable over the network (AV:N) with low complexity (AC:L) and no attack requirements (AT:N). It requires low privileges (PR:L) as any authenticated user can trigger the issue. No user interaction is required (UI:N). The impact is high on integrity (VI:H) due to unauthorized disclosure of private case data, but no impact on confidentiality (VC:N) or availability (VA:N) of the primary system, and no secondary impacts (SC:N, SI:N, SA:N).

Weakness rationale
  • CWE-639 The vulnerability allows access to resources (private case data) based on user-controlled inputs (task assignment or organization membership) without verifying the authorization context (case visibility). This is a classic authorization bypass where the application fails to enforce proper access controls on the parent resource.
Attack pattern rationale
  • CAPEC-126 The vulnerability allows users to access private case data by interacting with calendar endpoints that do not properly enforce case-level authorization. While not a direct URL manipulation, the effect is similar to forced browsing where the application fails to restrict access to resources based on the user's privileges. The uncertainty lies in the specific mechanism, but the impact is unauthorized access to restricted data.
Assumptions to verify
  • The CAPEC mapping is uncertain as the specific attack pattern is not explicitly detailed in the patch, but CAPEC-126 is the closest plausible match for unauthorized access to restricted data.
  • The CVSS vector assumes that the primary impact is on integrity (VI:H) due to unauthorized disclosure, but confidentiality (VC:N) is marked as none based on the assumption that the data is not considered highly sensitive in this context. This may need adjustment based on the sensitivity of the case data.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 10 11 medium 2
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-17 08:41 UTC Fix developed Corrective change authored (98196d0523d4e1136014574ff0b526da1d92381b): fix: [security] Recheck private case access in task calendars https://github.com/flowintel/flowintel/commit/98196d0523d4e1136014574ff0b526da1d92381b
  2. 2026-11-20 14:19 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/98196d0523d4e1136014574ff0b526da1d92381b

{
  "containers": {
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/calendar/calendar_core.py"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "A vulnerability exists in the flowintel application\u0027s calendar functionality that allows unauthorized access to private case data. The task calendar queries failed to enforce visibility checks on the parent case associated with tasks. Specifically, tasks assigned to a user or belonging to the user\u0027s organization were returned in calendar views (including browser and API ICS feeds) without verifying whether the user had permission to view the parent case. This allowed users to potentially discover and access details of private cases they were not authorized to view, leading to information disclosure.\n\nThe issue affected the default task visibility logic and the assigned-only task queries. The fix ensures that case visibility is rechecked in addition to task assignment or organization membership, preventing the exposure of private case data in calendar outputs.\n\nAffected versions: flowintel \u003c 4.1.0"
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-126",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-126 Forced Browsing"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 7.1,
            "baseSeverity": "HIGH",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:H/VA:N/SC:N/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "HIGH",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T14:19:51Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-639",
              "description": "CWE-639 Authorization Bypass Through User-Controlled Key",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "orgId": "00000000-0000-4000-8000-000000000000"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/98196d0523d4e1136014574ff0b526da1d92381b"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The vulnerability is remediated by adding explicit case visibility checks to the task calendar queries. The fix ensures that tasks are only visible if the user has permission to view the parent case, in addition to being assigned to the task or belonging to the same organization. This prevents the disclosure of private case data in calendar views and ICS feeds."
        }
      ],
      "title": "flowintel: Insecure Direct Object Reference in Task Calendar Access Control",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The CAPEC mapping is uncertain as the specific attack pattern is not explicitly detailed in the patch, but CAPEC-126 is the closest plausible match for unauthorized access to restricted data.",
                  "The CVSS vector assumes that the primary impact is on integrity (VI:H) due to unauthorized disclosure, but confidentiality (VC:N) is marked as none based on the assumption that the data is not considered highly sensitive in this context. This may need adjustment based on the sensitivity of the case data."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-126",
                    "rationale": "The vulnerability allows users to access private case data by interacting with calendar endpoints that do not properly enforce case-level authorization. While not a direct URL manipulation, the effect is similar to forced browsing where the application fails to restrict access to resources based on the user\u0027s privileges. The uncertainty lies in the specific mechanism, but the impact is unauthorized access to restricted data."
                  }
                ],
                "commit": "98196d0523d4e1136014574ff0b526da1d92381b",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "The vulnerability is exploitable over the network (AV:N) with low complexity (AC:L) and no attack requirements (AT:N). It requires low privileges (PR:L) as any authenticated user can trigger the issue. No user interaction is required (UI:N). The impact is high on integrity (VI:H) due to unauthorized disclosure of private case data, but no impact on confidentiality (VC:N) or availability (VA:N) of the primary system, and no secondary impacts (SC:N, SI:N, SA:N).",
                "fixSummary": "The vulnerability is remediated by adding explicit case visibility checks to the task calendar queries. The fix ensures that tasks are only visible if the user has permission to view the parent case, in addition to being assigned to the task or belonging to the same organization. This prevents the disclosure of private case data in calendar views and ICS feeds.",
                "generatedAt": "2026-11-20T14:19:51.561363Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 2,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 10
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "618096dd62335667906e545842f94c0eadf5d9d9732f43bc5873d61306321c61",
                "patchSummary": "The patch modifies the `_tasks_visible_to_user_query` and `_tasks_assigned_to_user_query` functions in `app/calendar/calendar_core.py`. It replaces the previous outer join and filter logic with subqueries for task assignment and organization membership. A new filter condition is added to ensure that the case is not private or that the user has membership in the case\u0027s organization. This enforces case-level access control in calendar task queries.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "98196d0523d4e1136014574ff0b526da1d92381b",
                    "date": "Tue, 17 Nov 2026 09:41:44 +0100",
                    "patchSha256": "618096dd62335667906e545842f94c0eadf5d9d9732f43bc5873d61306321c61",
                    "retrievedAt": "2026-11-20T14:19:18.380521Z",
                    "source": "https://github.com/flowintel/flowintel/commit/98196d0523d4e1136014574ff0b526da1d92381b",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/98196d0523d4e1136014574ff0b526da1d92381b",
                    "subject": "fix: [security] Recheck private case access in task calendars"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/98196d0523d4e1136014574ff0b526da1d92381b",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T14:19:51Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Recheck private case access in task calendars",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-639",
                    "rationale": "The vulnerability allows access to resources (private case data) based on user-controlled inputs (task assignment or organization membership) without verifying the authorization context (case visibility). This is a classic authorization bypass where the application fails to enforce proper access controls on the parent resource."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (98196d0523d4e1136014574ff0b526da1d92381b): fix: [security] Recheck private case access in task calendars",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/98196d0523d4e1136014574ff0b526da1d92381b"
                    ],
                    "timestamp": "2026-11-17T08:41:44Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/98196d0523d4e1136014574ff0b526da1d92381b"
                    ],
                    "timestamp": "2026-11-20T14:19:18.380521Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20417"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "00000000-0000-4000-8000-000000000000",
    "cveId": "CVE-2026-133849",
    "datePublished": "2026-11-20T14:35:56.911289Z",
    "dateReserved": "2026-11-20T14:39:25.925Z",
    "dateUpdated": "2026-11-20T14:39:27.499705Z",
    "state": "PUBLISHED",
    "vulnId": "GCVE-1-2026-20417"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

GCVE-1-2026-20416 (CVE-2026-133848)

Vulnerability from gna-1 – Published: 2026-11-20 14:35 – Updated: 2026-11-20 14:39
VLAI
Title
flowintel: Organization Admin Can Delete Platform Admin Accounts Due to Missing Privilege Hierarchy Check
Summary
flowintel contains an authorization flaw in its user deletion functionality. Organization-scoped administrators (org admins) are able to delete platform-level administrator accounts through both the browser-based admin interface and the REST API. The existing authorization logic verified that an org admin could only target users within their own organization, but it did not enforce the privilege hierarchy between organization admins and platform admins. As a result, an authenticated org admin could issue a deletion request against any platform-admin account belonging to the same organization, causing the target account, its sessions, and related records to be permanently removed. Impact: - Loss of platform-level administrative oversight - Potential disruption of platform-wide management functions - Removal of higher-privilege accounts that should be protected from lower-privilege roles Affected versions: flowintel prior to 4.1.0
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CNA (v2.0.3)
Decision recorded 2026-11-20 14:19 UTC
CWE
Assigner
CIRCL GNA GNA-1
GNA scorecard A 92/100 over 319 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 14:19
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb 05b89a0d2175…
Confidence
medium
Commit Subject Patch SHA-256
e3d0b2ab153d fix: [security] Prevent organization admins from deleting 05b89a0d2175…
Fix summary

The fix introduces an explicit privilege-hierarchy check in both the browser and API deletion routes. Before any session invalidation or account deletion is performed, the system now verifies whether the target user holds platform-admin capability. If the requester is a pure organization admin and the target is a platform admin, the request is rejected with a 403 response, preserving the intended authorization boundary while leaving legitimate deletion workflows unaffected.

Patch summary

Two lines added in app/admin/admin.py (delete_user function) and two lines added in app/admin/admin_api.py (API GET handler). In each location, after the existing same-organization check for pure org admins, a new condition checks user.is_admin() / user_to_delete.is_admin(). If true, a 403 JSON response is returned with an appropriate error message, preventing the deletion from proceeding. No other logic was modified.

CVSS rationale

AV:N - the vulnerability is exploitable over the network via the web UI or REST API. AC:L - exploitation requires only a standard authenticated request targeting a platform-admin account; no race conditions or special timing. AT:N - no attack-target manipulation is needed. PR:L - the attacker must hold a valid organization-admin account. UI:N - no victim interaction is required; the org admin acts on their own behalf. VC:N - no confidential data is exposed. VI:N - the application's own data integrity is not corrupted; the deletion is a legitimate operation performed by an authenticated user. VA:H - platform-admin accounts and their sessions are permanently deleted, significantly degrading the availability of platform administration. SC:N - no cross-component confidentiality impact. SI:N - no cross-component integrity impact. SA:H - the broader platform administration and oversight functions lose availability when platform-admin accounts are removed.

Weakness rationale
  • CWE-285 The authorization logic for the user-deletion endpoint verified organizational scope but failed to enforce the privilege hierarchy between organization admins and platform admins. The missing check allowed a lower-privilege role to perform an action (deleting a higher-privilege account) that should have been restricted, constituting an improper authorization decision.
Attack pattern rationale
  • CAPEC-126 An authenticated organization admin leverages a missing authorization boundary to perform an action (deleting a platform-admin account) that is reserved for a higher-privilege role. This is a vertical privilege escalation: the actor gains the effective capability of a platform admin by removing platform-admin accounts. The mapping is the closest standard CAPEC; the attack does not involve traditional credential theft or token manipulation but rather an authorization logic gap that permits a lower-privilege actor to act at a higher privilege level.
Assumptions to verify
  • The affected version range is inferred from the explicit context (fixed in 4.1.0); the exact earliest affected version is not stated in the patch and is recorded as unspecified.
  • The CAPEC-126 mapping is the closest standard pattern; the vulnerability is more precisely an authorization-logic gap than a classic privilege-escalation exploit, but no more specific CAPEC exists for missing privilege-hierarchy checks in multi-tenant admin systems.
  • CVSS VA:H and SA:H assume that deleting platform-admin accounts materially degrades the availability of platform-level administration; if the deployment has redundant platform admins, the practical impact may be lower.
  • The patch metadata lists affected_version as '0', which is treated as a placeholder; the actual lower bound of affected versions is unknown.
  • No evidence of active exploitation or public PoC was found in the patch or metadata.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 7 11 medium 5
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-17 08:09 UTC Fix developed Corrective change authored (e3d0b2ab153d74df72a1f0c417dc2eef74554bcb): fix: [security] Prevent organization admins from deleting https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb
  2. 2026-11-20 14:18 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb

{
  "containers": {
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/admin/admin.py",
            "app/admin/admin_api.py"
          ],
          "product": "flowintel",
          "programFiles": [
            "app/admin/admin.py",
            "app/admin/admin_api.py"
          ],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "flowintel contains an authorization flaw in its user deletion functionality. Organization-scoped administrators (org admins) are able to delete platform-level administrator accounts through both the browser-based admin interface and the REST API.\n\nThe existing authorization logic verified that an org admin could only target users within their own organization, but it did not enforce the privilege hierarchy between organization admins and platform admins. As a result, an authenticated org admin could issue a deletion request against any platform-admin account belonging to the same organization, causing the target account, its sessions, and related records to be permanently removed.\n\nImpact:\n- Loss of platform-level administrative oversight\n- Potential disruption of platform-wide management functions\n- Removal of higher-privilege accounts that should be protected from lower-privilege roles\n\nAffected versions: flowintel prior to 4.1.0"
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-126",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-126 Privilege Escalation"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 8.3,
            "baseSeverity": "HIGH",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "HIGH",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:H",
            "version": "4.0",
            "vulnAvailabilityImpact": "HIGH",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "NONE",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T14:19:16Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-285",
              "description": "CWE-285 Improper Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "orgId": "00000000-0000-4000-8000-000000000000"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The fix introduces an explicit privilege-hierarchy check in both the browser and API deletion routes. Before any session invalidation or account deletion is performed, the system now verifies whether the target user holds platform-admin capability. If the requester is a pure organization admin and the target is a platform admin, the request is rejected with a 403 response, preserving the intended authorization boundary while leaving legitimate deletion workflows unaffected."
        }
      ],
      "title": "flowintel: Organization Admin Can Delete Platform Admin Accounts Due to Missing Privilege Hierarchy Check",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected version range is inferred from the explicit context (fixed in 4.1.0); the exact earliest affected version is not stated in the patch and is recorded as unspecified.",
                  "The CAPEC-126 mapping is the closest standard pattern; the vulnerability is more precisely an authorization-logic gap than a classic privilege-escalation exploit, but no more specific CAPEC exists for missing privilege-hierarchy checks in multi-tenant admin systems.",
                  "CVSS VA:H and SA:H assume that deleting platform-admin accounts materially degrades the availability of platform-level administration; if the deployment has redundant platform admins, the practical impact may be lower.",
                  "The patch metadata lists affected_version as \u00270\u0027, which is treated as a placeholder; the actual lower bound of affected versions is unknown.",
                  "No evidence of active exploitation or public PoC was found in the patch or metadata."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-126",
                    "rationale": "An authenticated organization admin leverages a missing authorization boundary to perform an action (deleting a platform-admin account) that is reserved for a higher-privilege role. This is a vertical privilege escalation: the actor gains the effective capability of a platform admin by removing platform-admin accounts. The mapping is the closest standard CAPEC; the attack does not involve traditional credential theft or token manipulation but rather an authorization logic gap that permits a lower-privilege actor to act at a higher privilege level."
                  }
                ],
                "commit": "e3d0b2ab153d74df72a1f0c417dc2eef74554bcb",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N - the vulnerability is exploitable over the network via the web UI or REST API. AC:L - exploitation requires only a standard authenticated request targeting a platform-admin account; no race conditions or special timing. AT:N - no attack-target manipulation is needed. PR:L - the attacker must hold a valid organization-admin account. UI:N - no victim interaction is required; the org admin acts on their own behalf. VC:N - no confidential data is exposed. VI:N - the application\u0027s own data integrity is not corrupted; the deletion is a legitimate operation performed by an authenticated user. VA:H - platform-admin accounts and their sessions are permanently deleted, significantly degrading the availability of platform administration. SC:N - no cross-component confidentiality impact. SI:N - no cross-component integrity impact. SA:H - the broader platform administration and oversight functions lose availability when platform-admin accounts are removed.",
                "fixSummary": "The fix introduces an explicit privilege-hierarchy check in both the browser and API deletion routes. Before any session invalidation or account deletion is performed, the system now verifies whether the target user holds platform-admin capability. If the requester is a pure organization admin and the target is a platform admin, the request is rejected with a 403 response, preserving the intended authorization boundary while leaving legitimate deletion workflows unaffected.",
                "generatedAt": "2026-11-20T14:19:16.357167Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 5,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 7
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "05b89a0d21756a6aeff8a9274a8c1993f718a0d5253b3aeabfe7965dfe04c470",
                "patchSummary": "Two lines added in app/admin/admin.py (delete_user function) and two lines added in app/admin/admin_api.py (API GET handler). In each location, after the existing same-organization check for pure org admins, a new condition checks user.is_admin() / user_to_delete.is_admin(). If true, a 403 JSON response is returned with an appropriate error message, preventing the deletion from proceeding. No other logic was modified.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "e3d0b2ab153d74df72a1f0c417dc2eef74554bcb",
                    "date": "Tue, 17 Nov 2026 09:09:12 +0100",
                    "patchSha256": "05b89a0d21756a6aeff8a9274a8c1993f718a0d5253b3aeabfe7965dfe04c470",
                    "retrievedAt": "2026-11-20T14:18:48.516621Z",
                    "source": "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb",
                    "subject": "fix: [security] Prevent organization admins from deleting"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T14:19:16Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Prevent organization admins from deleting",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-285",
                    "rationale": "The authorization logic for the user-deletion endpoint verified organizational scope but failed to enforce the privilege hierarchy between organization admins and platform admins. The missing check allowed a lower-privilege role to perform an action (deleting a higher-privilege account) that should have been restricted, constituting an improper authorization decision."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (e3d0b2ab153d74df72a1f0c417dc2eef74554bcb): fix: [security] Prevent organization admins from deleting",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb"
                    ],
                    "timestamp": "2026-11-17T08:09:12Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb"
                    ],
                    "timestamp": "2026-11-20T14:18:48.516621Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20416"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "00000000-0000-4000-8000-000000000000",
    "cveId": "CVE-2026-133848",
    "datePublished": "2026-11-20T14:35:48.426115Z",
    "dateReserved": "2026-11-20T14:39:06.507Z",
    "dateUpdated": "2026-11-20T14:39:07.954298Z",
    "state": "PUBLISHED",
    "vulnId": "GCVE-1-2026-20416"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

GCVE-1-2026-20415 (CVE-2026-133847)

Vulnerability from gna-1 – Published: 2026-11-20 14:35 – Updated: 2026-11-20 14:38
VLAI
Title
flowintel: Missing Authorization Check Allows Organization Admin to Edit Platform Admin Accounts via Browser Interface
Summary
flowintel contains an authorization bypass in the browser-based admin user editing interface. The edit_user view in the admin panel verified that a non-platform-admin user could only edit accounts belonging to their own organization, but it did not verify that the target account was not a platform-admin account. As a result, an authenticated user holding an organization-admin role could open the edit form for any platform-admin account and submit modifications, including changes to credentials, roles, or other account attributes. The API layer already enforced a restriction preventing non-admin users from targeting admin accounts, but the browser-based admin view lacked the equivalent check, creating an inconsistent authorization boundary. Impact: - Unauthorized modification of platform-admin account data (integrity) - Potential lockout or privilege manipulation of platform administrators (availability) - Escalation of an organization-admin role to effectively control platform-level administration Affected versions: flowintel prior to 4.1.0.
SSVC
Exploitation: none Automatable: yes Technical Impact: total
Supplier · CNA (v2.0.3)
Decision recorded 2026-11-20 14:23 UTC
CWE
Assigner
CIRCL GNA GNA-1
GNA scorecard A 92/100 over 319 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 14:23
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/db694be89ed47f203937ab97c6fe4b93c1085a90 2fdb8a47ac79…
Confidence
medium
Commit Subject Patch SHA-256
db694be89ed4 fix: [security] Prevent organization admins from editing 2fdb8a47ac79…
Fix summary

The fix adds an explicit authorization check in the browser-based user edit view that rejects the request before rendering or processing the form when the target account is a platform-admin account and the current user is not a platform administrator. This aligns the browser interface with the existing API-level restriction, ensuring that only platform administrators can modify platform-admin accounts.

Patch summary

Three lines are inserted in the edit_user function in app/admin/admin.py, immediately after the same-organization check and before the non-admin permission check. The added code calls user_modif.is_admin() on the target user; if true, it flashes an error message ('Only administrators can edit Admin users.') and redirects to the admin users list, preventing the form from being rendered or processed. No other files or logic are modified.

CVSS rationale

AV:N: web-based admin interface accessible over the network. AC:L: the attacker simply navigates to the edit URL for a known admin account; no race conditions or special timing required. AT:N: no manipulation of a separate attack target. PR:L: requires an authenticated account with organization-admin privileges. UI:N: no victim interaction beyond the attacker's own browser session. VC:N: the primary impact is modification, not disclosure; the attacker already has visibility into the user list. VI:H: the attacker can alter platform-admin credentials, roles, or settings. VA:H: the attacker can disable or lock out platform administrators. SC:N: no impact on separate components. SI:H: integrity of admin account records is compromised. SA:N: no impact on the security authority of other components.

Weakness rationale
  • CWE-862 The edit_user view was missing an authorization check that would prevent a non-platform-admin user from targeting a platform-admin account. The check existed in the API layer but was absent in the browser-based view, constituting a missing authorization control for a specific operation.
Attack pattern rationale
  • CAPEC-100 The attack exploits an incomplete authorization logic in the admin user-edit view: the same-organization check passes for a platform-admin account that happens to share the organization, but the missing admin-target check allows the lower-privileged user to proceed. This is a logic flaw in the access-control flow. CAPEC-100 is the closest available pattern; the exact CAPEC for a single missing check in a multi-step authorization flow is not explicitly defined in the CAPEC catalog, so this is the best plausible match.
Assumptions to verify
  • The affected version range is inferred from the explicit context (affected_version: 0, fixed_version: 4.1.0); the exact starting version is not independently confirmed by the patch alone.
  • The assumption that the API layer already enforced the admin-target restriction is taken from the commit message; the API code is not included in the patch for independent verification.
  • CAPEC-100 is selected as the closest plausible attack pattern; no CAPEC entry precisely describes a single missing authorization check in a multi-step view function, so this mapping carries moderate uncertainty.
  • CVSS VC is set to None on the assumption that the attacker already has visibility into the user list in the admin panel; if the edit form reveals additional sensitive fields not otherwise visible, VC could be Low.
  • The patch does not include the full edit_user function or surrounding context; the analysis assumes the three added lines are the complete remediation for this specific authorization gap.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 7 11 medium 5
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-17 07:44 UTC Fix developed Corrective change authored (db694be89ed47f203937ab97c6fe4b93c1085a90): fix: [security] Prevent organization admins from editing https://github.com/flowintel/flowintel/commit/db694be89ed47f203937ab97c6fe4b93c1085a90
  2. 2026-11-20 14:22 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/db694be89ed47f203937ab97c6fe4b93c1085a90

{
  "containers": {
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/admin/admin.py"
          ],
          "product": "flowintel",
          "programFiles": [
            "app/admin/admin.py"
          ],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "flowintel contains an authorization bypass in the browser-based admin user editing interface. The edit_user view in the admin panel verified that a non-platform-admin user could only edit accounts belonging to their own organization, but it did not verify that the target account was not a platform-admin account. As a result, an authenticated user holding an organization-admin role could open the edit form for any platform-admin account and submit modifications, including changes to credentials, roles, or other account attributes.\n\nThe API layer already enforced a restriction preventing non-admin users from targeting admin accounts, but the browser-based admin view lacked the equivalent check, creating an inconsistent authorization boundary.\n\nImpact:\n- Unauthorized modification of platform-admin account data (integrity)\n- Potential lockout or privilege manipulation of platform administrators (availability)\n- Escalation of an organization-admin role to effectively control platform-level administration\n\nAffected versions: flowintel prior to 4.1.0."
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-100",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-100 Leveraging Broken Logic to Modify Access Controls"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 8.4,
            "baseSeverity": "HIGH",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "HIGH",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:H/VA:H/SC:N/SI:H/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "HIGH",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "HIGH",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "total"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T14:23:03Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-862",
              "description": "CWE-862 Missing Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "orgId": "00000000-0000-4000-8000-000000000000"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/db694be89ed47f203937ab97c6fe4b93c1085a90"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The fix adds an explicit authorization check in the browser-based user edit view that rejects the request before rendering or processing the form when the target account is a platform-admin account and the current user is not a platform administrator. This aligns the browser interface with the existing API-level restriction, ensuring that only platform administrators can modify platform-admin accounts."
        }
      ],
      "title": "flowintel: Missing Authorization Check Allows Organization Admin to Edit Platform Admin Accounts via Browser Interface",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected version range is inferred from the explicit context (affected_version: 0, fixed_version: 4.1.0); the exact starting version is not independently confirmed by the patch alone.",
                  "The assumption that the API layer already enforced the admin-target restriction is taken from the commit message; the API code is not included in the patch for independent verification.",
                  "CAPEC-100 is selected as the closest plausible attack pattern; no CAPEC entry precisely describes a single missing authorization check in a multi-step view function, so this mapping carries moderate uncertainty.",
                  "CVSS VC is set to None on the assumption that the attacker already has visibility into the user list in the admin panel; if the edit form reveals additional sensitive fields not otherwise visible, VC could be Low.",
                  "The patch does not include the full edit_user function or surrounding context; the analysis assumes the three added lines are the complete remediation for this specific authorization gap."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-100",
                    "rationale": "The attack exploits an incomplete authorization logic in the admin user-edit view: the same-organization check passes for a platform-admin account that happens to share the organization, but the missing admin-target check allows the lower-privileged user to proceed. This is a logic flaw in the access-control flow. CAPEC-100 is the closest available pattern; the exact CAPEC for a single missing check in a multi-step authorization flow is not explicitly defined in the CAPEC catalog, so this is the best plausible match."
                  }
                ],
                "commit": "db694be89ed47f203937ab97c6fe4b93c1085a90",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N: web-based admin interface accessible over the network. AC:L: the attacker simply navigates to the edit URL for a known admin account; no race conditions or special timing required. AT:N: no manipulation of a separate attack target. PR:L: requires an authenticated account with organization-admin privileges. UI:N: no victim interaction beyond the attacker\u0027s own browser session. VC:N: the primary impact is modification, not disclosure; the attacker already has visibility into the user list. VI:H: the attacker can alter platform-admin credentials, roles, or settings. VA:H: the attacker can disable or lock out platform administrators. SC:N: no impact on separate components. SI:H: integrity of admin account records is compromised. SA:N: no impact on the security authority of other components.",
                "fixSummary": "The fix adds an explicit authorization check in the browser-based user edit view that rejects the request before rendering or processing the form when the target account is a platform-admin account and the current user is not a platform administrator. This aligns the browser interface with the existing API-level restriction, ensuring that only platform administrators can modify platform-admin accounts.",
                "generatedAt": "2026-11-20T14:23:03.702620Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 5,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 7
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "2fdb8a47ac795b3fdce91f16c7cbffd444a7dcefb5a42ea0cfff7a8aa3922e38",
                "patchSummary": "Three lines are inserted in the edit_user function in app/admin/admin.py, immediately after the same-organization check and before the non-admin permission check. The added code calls user_modif.is_admin() on the target user; if true, it flashes an error message (\u0027Only administrators can edit Admin users.\u0027) and redirects to the admin users list, preventing the form from being rendered or processed. No other files or logic are modified.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "db694be89ed47f203937ab97c6fe4b93c1085a90",
                    "date": "Tue, 17 Nov 2026 08:44:08 +0100",
                    "patchSha256": "2fdb8a47ac795b3fdce91f16c7cbffd444a7dcefb5a42ea0cfff7a8aa3922e38",
                    "retrievedAt": "2026-11-20T14:22:37.444372Z",
                    "source": "https://github.com/flowintel/flowintel/commit/db694be89ed47f203937ab97c6fe4b93c1085a90",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/db694be89ed47f203937ab97c6fe4b93c1085a90",
                    "subject": "fix: [security] Prevent organization admins from editing"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/db694be89ed47f203937ab97c6fe4b93c1085a90",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "total"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T14:23:03Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Prevent organization admins from editing",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-862",
                    "rationale": "The edit_user view was missing an authorization check that would prevent a non-platform-admin user from targeting a platform-admin account. The check existed in the API layer but was absent in the browser-based view, constituting a missing authorization control for a specific operation."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (db694be89ed47f203937ab97c6fe4b93c1085a90): fix: [security] Prevent organization admins from editing",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/db694be89ed47f203937ab97c6fe4b93c1085a90"
                    ],
                    "timestamp": "2026-11-17T07:44:08Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/db694be89ed47f203937ab97c6fe4b93c1085a90"
                    ],
                    "timestamp": "2026-11-20T14:22:37.444372Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20415"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "00000000-0000-4000-8000-000000000000",
    "cveId": "CVE-2026-133847",
    "datePublished": "2026-11-20T14:35:45.429018Z",
    "dateReserved": "2026-11-20T14:38:46.730Z",
    "dateUpdated": "2026-11-20T14:38:47.618782Z",
    "state": "PUBLISHED",
    "vulnId": "GCVE-1-2026-20415"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

GCVE-1-2026-20414 (CVE-2026-133846)

Vulnerability from gna-1 – Published: 2026-11-20 14:35 – Updated: 2026-11-20 14:38
VLAI
Title
FlowIntel Missing Authorization Check Allows Unauthorized Revival of Completed Privileged Cases via Task Reopening
Summary
FlowIntel contains an authorization bypass in its task management workflow. When a user reopens a completed task (by toggling its completion state or changing its status to a non-finished value) or creates a new task within a completed privileged case, the parent case is automatically revived (un-completed, status reset to Created, finish date cleared). The system failed to enforce the required Admin or Case Admin privilege check before performing this case revival side effect. A user with basic task-level permissions (e.g., editor, queuer, or queue admin) could trigger the revival of a completed privileged case through the browser interface or the REST API, without holding the elevated role required for case-level operations. This could disrupt case lifecycle management, clear completion metadata, and potentially expose the case to further unauthorized modifications. The vulnerability affects both the browser-based task completion/status-change endpoints and the corresponding API endpoints, as well as the task creation flow. Affected versions: FlowIntel prior to 4.1.0.
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CNA (v2.0.3)
Decision recorded 2026-11-20 14:22 UTC
CWE
Assigner
CIRCL GNA GNA-1
GNA scorecard A 92/100 over 319 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 14:22
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/7eb9fa76d46010548c5552c0a8b1d438d839226c f7e85e27691e…
Confidence
medium
Commit Subject Patch SHA-256
7eb9fa76d460 fix: [security] Enforce privileged case permissions on task f7e85e27691e…
Fix summary

The fix introduces explicit authorization checks (can_reopen_case and can_reopen_task) that require Admin or Case Admin privileges before a completed privileged case can be revived as a side effect of task reopening or task creation. Denial is enforced at the core model layer and at both the browser and API endpoint layers, with audit logging of denied attempts. Legitimate task creation is preserved, but the automatic case revival is suppressed for unauthorized users.

Patch summary

Added can_reopen_case() and can_reopen_task() methods to TaskCore.py that verify Admin or Case Admin role when the parent case is both completed and privileged. Inserted permission checks in complete_task() and create_task() in TaskCore.py. Added 403 denial responses with audit log entries in task.py (browser endpoints for task completion and status change) and task_api.py (API endpoints for task completion and status change). Added 134 lines of regression tests covering role boundaries, UUID aliases, status IDs, API identity, Matrix delegation, denial side effects, and legitimate reopening workflows.

CVSS rationale

AV:N - exploitable over the network via web UI or REST API. AC:L - the attack requires only a standard task completion toggle or status change request; no race conditions or special timing. AT:N - no manipulation of the target system is needed. PR:L - requires an authenticated user with basic task permissions (editor, queuer, or queue admin). UI:N - no victim interaction required. VC:N - no new confidential data is exposed; the case data was already accessible. VI:L - the case state is modified (completed flag cleared, status reset, finish date removed), which is a limited integrity impact on case lifecycle metadata. VA:N - no denial of service; the case remains accessible. SC/SI/SA:N - no secondary system impact identified.

Weakness rationale
  • CWE-862 The system failed to enforce an authorization check before allowing a completed privileged case to be revived as a side effect of task reopening or task creation. The permission check for case-level operations was entirely absent from the task-level code paths.
Attack pattern rationale
  • CAPEC-114 The attacker performs an action (task reopening or task creation) that they are partially authorized to perform, but the system fails to enforce the higher-level authorization required for the side effect (privileged case revival). This is closest to forced browsing in the sense that the user effectively triggers an unauthorized state change on a resource (the case) they do not have full authority over. Uncertainty: the user IS authorized to interact with the task itself; the missing check is on the implicit case-level side effect, making this a partial rather than pure forced-browsing scenario.
Assumptions to verify
  • The affected version range is inferred from the explicit context (affected_version 0, fixed_version 4.1.0); the exact starting version of the vulnerable code is not confirmed by the patch alone.
  • CAPEC-114 (Forced Browsing) is the closest available mapping; the actual weakness is a missing authorization check on a side effect of an authorized action, which does not map precisely to any single CAPEC entry.
  • CVSS VI:L assumes the case state modification (revival) constitutes a limited integrity impact; if downstream workflows depend on case completion state, the practical impact could be higher.
  • The patch references a 'Matrix delegation' mechanism for API identity; the exact trust model for delegated requests is not fully documented in the patch.
  • PR:L assumes the attacker needs at minimum an editor-level account; the patch tests confirm editor, queuer, and queue_admin roles are affected.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 7 11 medium 5
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-16 08:28 UTC Fix developed Corrective change authored (7eb9fa76d46010548c5552c0a8b1d438d839226c): fix: [security] Enforce privileged case permissions on task https://github.com/flowintel/flowintel/commit/7eb9fa76d46010548c5552c0a8b1d438d839226c
  2. 2026-11-20 14:22 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/7eb9fa76d46010548c5552c0a8b1d438d839226c

{
  "containers": {
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/case/TaskCore.py",
            "app/case/task.py",
            "app/case/task_api.py"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "FlowIntel contains an authorization bypass in its task management workflow. When a user reopens a completed task (by toggling its completion state or changing its status to a non-finished value) or creates a new task within a completed privileged case, the parent case is automatically revived (un-completed, status reset to Created, finish date cleared). The system failed to enforce the required Admin or Case Admin privilege check before performing this case revival side effect.\n\nA user with basic task-level permissions (e.g., editor, queuer, or queue admin) could trigger the revival of a completed privileged case through the browser interface or the REST API, without holding the elevated role required for case-level operations. This could disrupt case lifecycle management, clear completion metadata, and potentially expose the case to further unauthorized modifications.\n\nThe vulnerability affects both the browser-based task completion/status-change endpoints and the corresponding API endpoints, as well as the task creation flow.\n\nAffected versions: FlowIntel prior to 4.1.0."
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-114",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-114 Forced Browsing"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 5.3,
            "baseSeverity": "MEDIUM",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:L/VA:N/SC:N/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "LOW",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T14:22:36Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-862",
              "description": "CWE-862 Missing Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "orgId": "00000000-0000-4000-8000-000000000000"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/7eb9fa76d46010548c5552c0a8b1d438d839226c"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The fix introduces explicit authorization checks (can_reopen_case and can_reopen_task) that require Admin or Case Admin privileges before a completed privileged case can be revived as a side effect of task reopening or task creation. Denial is enforced at the core model layer and at both the browser and API endpoint layers, with audit logging of denied attempts. Legitimate task creation is preserved, but the automatic case revival is suppressed for unauthorized users."
        }
      ],
      "title": "FlowIntel Missing Authorization Check Allows Unauthorized Revival of Completed Privileged Cases via Task Reopening",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected version range is inferred from the explicit context (affected_version 0, fixed_version 4.1.0); the exact starting version of the vulnerable code is not confirmed by the patch alone.",
                  "CAPEC-114 (Forced Browsing) is the closest available mapping; the actual weakness is a missing authorization check on a side effect of an authorized action, which does not map precisely to any single CAPEC entry.",
                  "CVSS VI:L assumes the case state modification (revival) constitutes a limited integrity impact; if downstream workflows depend on case completion state, the practical impact could be higher.",
                  "The patch references a \u0027Matrix delegation\u0027 mechanism for API identity; the exact trust model for delegated requests is not fully documented in the patch.",
                  "PR:L assumes the attacker needs at minimum an editor-level account; the patch tests confirm editor, queuer, and queue_admin roles are affected."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-114",
                    "rationale": "The attacker performs an action (task reopening or task creation) that they are partially authorized to perform, but the system fails to enforce the higher-level authorization required for the side effect (privileged case revival). This is closest to forced browsing in the sense that the user effectively triggers an unauthorized state change on a resource (the case) they do not have full authority over. Uncertainty: the user IS authorized to interact with the task itself; the missing check is on the implicit case-level side effect, making this a partial rather than pure forced-browsing scenario."
                  }
                ],
                "commit": "7eb9fa76d46010548c5552c0a8b1d438d839226c",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N - exploitable over the network via web UI or REST API. AC:L - the attack requires only a standard task completion toggle or status change request; no race conditions or special timing. AT:N - no manipulation of the target system is needed. PR:L - requires an authenticated user with basic task permissions (editor, queuer, or queue admin). UI:N - no victim interaction required. VC:N - no new confidential data is exposed; the case data was already accessible. VI:L - the case state is modified (completed flag cleared, status reset, finish date removed), which is a limited integrity impact on case lifecycle metadata. VA:N - no denial of service; the case remains accessible. SC/SI/SA:N - no secondary system impact identified.",
                "fixSummary": "The fix introduces explicit authorization checks (can_reopen_case and can_reopen_task) that require Admin or Case Admin privileges before a completed privileged case can be revived as a side effect of task reopening or task creation. Denial is enforced at the core model layer and at both the browser and API endpoint layers, with audit logging of denied attempts. Legitimate task creation is preserved, but the automatic case revival is suppressed for unauthorized users.",
                "generatedAt": "2026-11-20T14:22:36.914577Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 5,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 7
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "f7e85e27691e19d219a953bd0234e210ec548def95dfc215cc50682799be99da",
                "patchSummary": "Added can_reopen_case() and can_reopen_task() methods to TaskCore.py that verify Admin or Case Admin role when the parent case is both completed and privileged. Inserted permission checks in complete_task() and create_task() in TaskCore.py. Added 403 denial responses with audit log entries in task.py (browser endpoints for task completion and status change) and task_api.py (API endpoints for task completion and status change). Added 134 lines of regression tests covering role boundaries, UUID aliases, status IDs, API identity, Matrix delegation, denial side effects, and legitimate reopening workflows.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "7eb9fa76d46010548c5552c0a8b1d438d839226c",
                    "date": "Mon, 16 Nov 2026 09:28:51 +0100",
                    "patchSha256": "f7e85e27691e19d219a953bd0234e210ec548def95dfc215cc50682799be99da",
                    "retrievedAt": "2026-11-20T14:22:03.456534Z",
                    "source": "https://github.com/flowintel/flowintel/commit/7eb9fa76d46010548c5552c0a8b1d438d839226c",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/7eb9fa76d46010548c5552c0a8b1d438d839226c",
                    "subject": "fix: [security] Enforce privileged case permissions on task"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/7eb9fa76d46010548c5552c0a8b1d438d839226c",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T14:22:36Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Enforce privileged case permissions on task",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-862",
                    "rationale": "The system failed to enforce an authorization check before allowing a completed privileged case to be revived as a side effect of task reopening or task creation. The permission check for case-level operations was entirely absent from the task-level code paths."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (7eb9fa76d46010548c5552c0a8b1d438d839226c): fix: [security] Enforce privileged case permissions on task",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/7eb9fa76d46010548c5552c0a8b1d438d839226c"
                    ],
                    "timestamp": "2026-11-16T08:28:51Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/7eb9fa76d46010548c5552c0a8b1d438d839226c"
                    ],
                    "timestamp": "2026-11-20T14:22:03.456534Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20414"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "00000000-0000-4000-8000-000000000000",
    "cveId": "CVE-2026-133846",
    "datePublished": "2026-11-20T14:35:40.246147Z",
    "dateReserved": "2026-11-20T14:38:25.137Z",
    "dateUpdated": "2026-11-20T14:38:25.222531Z",
    "state": "PUBLISHED",
    "vulnId": "GCVE-1-2026-20414"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}
displaying 1 - 10 publications in total 395

CVE-2026-133848 (GCVE-0-2026-133848)

Vulnerability from cvelistv5 – Published: 2026-11-20 14:39 – Updated: 2026-11-20 18:31
VLAI
Title
flowintel: Organization Admin Can Delete Platform Admin Accounts Due to Missing Privilege Hierarchy Check
Summary
flowintel contains an authorization flaw in its user deletion functionality. Organization-scoped administrators (org admins) are able to delete platform-level administrator accounts through both the browser-based admin interface and the REST API. The existing authorization logic verified that an org admin could only target users within their own organization, but it did not enforce the privilege hierarchy between organization admins and platform admins. As a result, an authenticated org admin could issue a deletion request against any platform-admin account belonging to the same organization, causing the target account, its sessions, and related records to be permanently removed. Impact: - Loss of platform-level administrative oversight - Potential disruption of platform-wide management functions - Removal of higher-privilege accounts that should be protected from lower-privilege roles Affected versions: flowintel prior to 4.1.0
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 14:19 UTC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:31 UTC
CWE
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 14:19
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb 05b89a0d2175…
Confidence
medium
Commit Subject Patch SHA-256
e3d0b2ab153d fix: [security] Prevent organization admins from deleting 05b89a0d2175…
Fix summary

The fix introduces an explicit privilege-hierarchy check in both the browser and API deletion routes. Before any session invalidation or account deletion is performed, the system now verifies whether the target user holds platform-admin capability. If the requester is a pure organization admin and the target is a platform admin, the request is rejected with a 403 response, preserving the intended authorization boundary while leaving legitimate deletion workflows unaffected.

Patch summary

Two lines added in app/admin/admin.py (delete_user function) and two lines added in app/admin/admin_api.py (API GET handler). In each location, after the existing same-organization check for pure org admins, a new condition checks user.is_admin() / user_to_delete.is_admin(). If true, a 403 JSON response is returned with an appropriate error message, preventing the deletion from proceeding. No other logic was modified.

CVSS rationale

AV:N - the vulnerability is exploitable over the network via the web UI or REST API. AC:L - exploitation requires only a standard authenticated request targeting a platform-admin account; no race conditions or special timing. AT:N - no attack-target manipulation is needed. PR:L - the attacker must hold a valid organization-admin account. UI:N - no victim interaction is required; the org admin acts on their own behalf. VC:N - no confidential data is exposed. VI:N - the application's own data integrity is not corrupted; the deletion is a legitimate operation performed by an authenticated user. VA:H - platform-admin accounts and their sessions are permanently deleted, significantly degrading the availability of platform administration. SC:N - no cross-component confidentiality impact. SI:N - no cross-component integrity impact. SA:H - the broader platform administration and oversight functions lose availability when platform-admin accounts are removed.

Weakness rationale
  • CWE-285 The authorization logic for the user-deletion endpoint verified organizational scope but failed to enforce the privilege hierarchy between organization admins and platform admins. The missing check allowed a lower-privilege role to perform an action (deleting a higher-privilege account) that should have been restricted, constituting an improper authorization decision.
Attack pattern rationale
  • CAPEC-126 An authenticated organization admin leverages a missing authorization boundary to perform an action (deleting a platform-admin account) that is reserved for a higher-privilege role. This is a vertical privilege escalation: the actor gains the effective capability of a platform admin by removing platform-admin accounts. The mapping is the closest standard CAPEC; the attack does not involve traditional credential theft or token manipulation but rather an authorization logic gap that permits a lower-privilege actor to act at a higher privilege level.
Assumptions to verify
  • The affected version range is inferred from the explicit context (fixed in 4.1.0); the exact earliest affected version is not stated in the patch and is recorded as unspecified.
  • The CAPEC-126 mapping is the closest standard pattern; the vulnerability is more precisely an authorization-logic gap than a classic privilege-escalation exploit, but no more specific CAPEC exists for missing privilege-hierarchy checks in multi-tenant admin systems.
  • CVSS VA:H and SA:H assume that deleting platform-admin accounts materially degrades the availability of platform-level administration; if the deployment has redundant platform admins, the practical impact may be lower.
  • The patch metadata lists affected_version as '0', which is treated as a placeholder; the actual lower bound of affected versions is unknown.
  • No evidence of active exploitation or public PoC was found in the patch or metadata.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 7 11 medium 5
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-17 08:09 UTC Fix developed Corrective change authored (e3d0b2ab153d74df72a1f0c417dc2eef74554bcb): fix: [security] Prevent organization admins from deleting https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb
  2. 2026-11-20 14:18 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133848",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:31:26.526789Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:31:42.592Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/admin/admin.py",
            "app/admin/admin_api.py"
          ],
          "product": "flowintel",
          "programFiles": [
            "app/admin/admin.py",
            "app/admin/admin_api.py"
          ],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "flowintel contains an authorization flaw in its user deletion functionality. Organization-scoped administrators (org admins) are able to delete platform-level administrator accounts through both the browser-based admin interface and the REST API.\n\nThe existing authorization logic verified that an org admin could only target users within their own organization, but it did not enforce the privilege hierarchy between organization admins and platform admins. As a result, an authenticated org admin could issue a deletion request against any platform-admin account belonging to the same organization, causing the target account, its sessions, and related records to be permanently removed.\n\nImpact:\n- Loss of platform-level administrative oversight\n- Potential disruption of platform-wide management functions\n- Removal of higher-privilege accounts that should be protected from lower-privilege roles\n\nAffected versions: flowintel prior to 4.1.0"
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-126",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-126 Privilege Escalation"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 8.3,
            "baseSeverity": "HIGH",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "HIGH",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:H",
            "version": "4.0",
            "vulnAvailabilityImpact": "HIGH",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "NONE",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T14:19:16Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-285",
              "description": "CWE-285 Improper Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T14:39:11.588Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The fix introduces an explicit privilege-hierarchy check in both the browser and API deletion routes. Before any session invalidation or account deletion is performed, the system now verifies whether the target user holds platform-admin capability. If the requester is a pure organization admin and the target is a platform admin, the request is rejected with a 403 response, preserving the intended authorization boundary while leaving legitimate deletion workflows unaffected."
        }
      ],
      "title": "flowintel: Organization Admin Can Delete Platform Admin Accounts Due to Missing Privilege Hierarchy Check",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected version range is inferred from the explicit context (fixed in 4.1.0); the exact earliest affected version is not stated in the patch and is recorded as unspecified.",
                  "The CAPEC-126 mapping is the closest standard pattern; the vulnerability is more precisely an authorization-logic gap than a classic privilege-escalation exploit, but no more specific CAPEC exists for missing privilege-hierarchy checks in multi-tenant admin systems.",
                  "CVSS VA:H and SA:H assume that deleting platform-admin accounts materially degrades the availability of platform-level administration; if the deployment has redundant platform admins, the practical impact may be lower.",
                  "The patch metadata lists affected_version as \u00270\u0027, which is treated as a placeholder; the actual lower bound of affected versions is unknown.",
                  "No evidence of active exploitation or public PoC was found in the patch or metadata."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-126",
                    "rationale": "An authenticated organization admin leverages a missing authorization boundary to perform an action (deleting a platform-admin account) that is reserved for a higher-privilege role. This is a vertical privilege escalation: the actor gains the effective capability of a platform admin by removing platform-admin accounts. The mapping is the closest standard CAPEC; the attack does not involve traditional credential theft or token manipulation but rather an authorization logic gap that permits a lower-privilege actor to act at a higher privilege level."
                  }
                ],
                "commit": "e3d0b2ab153d74df72a1f0c417dc2eef74554bcb",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N - the vulnerability is exploitable over the network via the web UI or REST API. AC:L - exploitation requires only a standard authenticated request targeting a platform-admin account; no race conditions or special timing. AT:N - no attack-target manipulation is needed. PR:L - the attacker must hold a valid organization-admin account. UI:N - no victim interaction is required; the org admin acts on their own behalf. VC:N - no confidential data is exposed. VI:N - the application\u0027s own data integrity is not corrupted; the deletion is a legitimate operation performed by an authenticated user. VA:H - platform-admin accounts and their sessions are permanently deleted, significantly degrading the availability of platform administration. SC:N - no cross-component confidentiality impact. SI:N - no cross-component integrity impact. SA:H - the broader platform administration and oversight functions lose availability when platform-admin accounts are removed.",
                "fixSummary": "The fix introduces an explicit privilege-hierarchy check in both the browser and API deletion routes. Before any session invalidation or account deletion is performed, the system now verifies whether the target user holds platform-admin capability. If the requester is a pure organization admin and the target is a platform admin, the request is rejected with a 403 response, preserving the intended authorization boundary while leaving legitimate deletion workflows unaffected.",
                "generatedAt": "2026-11-20T14:19:16.357167Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 5,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 7
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "05b89a0d21756a6aeff8a9274a8c1993f718a0d5253b3aeabfe7965dfe04c470",
                "patchSummary": "Two lines added in app/admin/admin.py (delete_user function) and two lines added in app/admin/admin_api.py (API GET handler). In each location, after the existing same-organization check for pure org admins, a new condition checks user.is_admin() / user_to_delete.is_admin(). If true, a 403 JSON response is returned with an appropriate error message, preventing the deletion from proceeding. No other logic was modified.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "e3d0b2ab153d74df72a1f0c417dc2eef74554bcb",
                    "date": "Tue, 17 Nov 2026 09:09:12 +0100",
                    "patchSha256": "05b89a0d21756a6aeff8a9274a8c1993f718a0d5253b3aeabfe7965dfe04c470",
                    "retrievedAt": "2026-11-20T14:18:48.516621Z",
                    "source": "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb",
                    "subject": "fix: [security] Prevent organization admins from deleting"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T14:19:16Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Prevent organization admins from deleting",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-285",
                    "rationale": "The authorization logic for the user-deletion endpoint verified organizational scope but failed to enforce the privilege hierarchy between organization admins and platform admins. The missing check allowed a lower-privilege role to perform an action (deleting a higher-privilege account) that should have been restricted, constituting an improper authorization decision."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (e3d0b2ab153d74df72a1f0c417dc2eef74554bcb): fix: [security] Prevent organization admins from deleting",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb"
                    ],
                    "timestamp": "2026-11-17T08:09:12Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/e3d0b2ab153d74df72a1f0c417dc2eef74554bcb"
                    ],
                    "timestamp": "2026-11-20T14:18:48.516621Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20416"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133848",
    "datePublished": "2026-11-20T14:39:11.588Z",
    "dateReserved": "2026-11-20T14:39:06.507Z",
    "dateUpdated": "2026-11-20T18:31:42.592Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133667 (GCVE-0-2026-133667)

Vulnerability from cvelistv5 – Published: 2026-11-20 10:58 – Updated: 2026-11-20 18:26
VLAI
Title
flowintel DOM-based XSS via Timeline Autolinking, Search Highlighting, and Content-Type Confusion
Summary
flowintel contains multiple related flaws that allow an authenticated attacker to inject arbitrary HTML or script into pages viewed by other users. - Timeline rendering: the timeline library's autolinking feature rewrote already-sanitised case notes and imported MISP attribute values into generated HTML link elements. Crafted text containing quote characters could break out of the generated attribute context, altering the surrounding markup and enabling script execution in the victim's browser. - Search highlighting: the highlight injection logic used a regular expression to locate text between angle brackets and spliced <mark> elements into the resulting string. Because the pattern could match inside HTML attribute values, the rewrite could corrupt or inject content into attributes of already-sanitised markup, again leading to DOM-based cross-site scripting. - Content-Type confusion: three analyser selection endpoints returned session-stored JSON strings with a text/html content type. A browser receiving such a response will parse and render the body as HTML, so any HTML or script embedded in the stored value is executed in the victim's context. Preconditions: the attacker must be an authenticated user with the ability to create or import case content (notes, MISP attributes) containing specially crafted quote or angle-bracket sequences. The victim must subsequently open the affected case timeline, run a search, or request one of the selection endpoints. Impact: successful exploitation yields DOM-based cross-site scripting in the victim's browser session, potentially allowing disclosure of case data, session manipulation, or execution of arbitrary actions under the victim's identity. Affected versions: flowintel prior to 4.1.0.
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 10:50 UTC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:26 UTC
CWE
  • CWE-79 - Improper Neutralization of Input in Web Page ('Cross-site Scripting')
  • CWE-799 - Improper Control of Resources Identifiers
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 10:50
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b.patch 5352fc62c6a6…
Confidence
medium
Commit Subject Patch SHA-256
5f745040d79d fix: [security] Timeline and search rendering preserve text 5352fc62c6a6…
Fix summary

The vulnerability is remediated by eliminating the three injection paths: (1) autolinking is disabled on every timeline event so the library no longer rewrites sanitised HTML; (2) search and auto-highlight marks are now created by walking the DOM text nodes and setting textContent, which cannot alter existing attributes or split character entities; (3) the three selection endpoints now decode the stored JSON once, respond with application/json and X-Content-Type-Options: nosniff, and reject malformed or excessively nested payloads with a 400 JSON error, preventing the browser from interpreting the body as HTML.

Patch summary

In case_history.js and timeline_view.js, an autolink: false property is added to each TimelineJS event object to suppress linkification. In smart-render.js, the previous regex-based string replacement (html.replace(/>([^<]*)</g, ...)) is replaced with a DOM TreeWalker that iterates text nodes, builds <mark> elements via document.createElement and textContent, and splices them into a DocumentFragment, preserving all existing attributes and entities. A new count_html_matches helper counts matches across text nodes. The plain-text fallback branches that performed raw regex replacement on escaped strings are removed. In misp_modules.py, a new _selected_json_response helper decodes the session value with json.loads, wraps it in jsonify, sets X-Content-Type-Options: nosniff, and returns a 400 JSON error on TypeError, ValueError, or RecursionError. The three get_*_selected endpoints are updated to call this helper instead of returning the raw session string.

CVSS rationale

AV:N – the application is a web service reachable over the network. AC:L – crafting quote-bearing text in case notes or MISP attributes is straightforward; no race or special timing is needed. AT:N – no active attack or prior compromise is required beyond normal authenticated use. PR:L – the attacker needs an authenticated account to create or import the malicious content (endpoints are login_required). UI:A – the victim must actively open the case timeline, run a search, or request a selection endpoint for the injected markup to execute. VC:L – XSS can expose case data, session tokens, or other sensitive information visible in the browser. VI:L – the attacker can modify displayed content or trigger actions in the victim's session. VA:N – no evidence of denial-of-service impact beyond the nested-JSON guard. SC:N – no cross-component or cross-origin impact is demonstrated. SI:L – other authenticated users viewing the same case could be affected by the stored malicious content. SA:N – no elevation of security authority is indicated.

Weakness rationale
  • CWE-79 All three sub-issues (autolinking rewriting sanitised HTML, regex-based highlight injection corrupting attributes, and serving JSON as text/html) result in attacker-controlled content being interpreted as HTML/JavaScript in the victim's browser, which is the defining characteristic of DOM-based XSS.
  • CWE-799 The content-type confusion (serving application/json data as text/html) is a secondary weakness that amplifies the XSS by causing the browser to parse the response body as HTML. This is a narrower, supporting CWE.
Attack pattern rationale
  • CAPEC-6 The primary exploitation path is client-side: attacker-controlled text (case notes, MISP values, search terms) is processed by JavaScript in the browser (TimelineJS autolinking, regex-based highlight injection) and the resulting DOM manipulation injects executable markup. The content-type confusion endpoint is a secondary vector but the dominant pattern is DOM-based script injection. CAPEC-6 is the closest standard attack pattern; the mapping is confident because the patch explicitly addresses client-side HTML rewriting of user-supplied content.
  • CAPEC-1 The three selection endpoints that returned session JSON as text/html could be leveraged as a reflected XSS vector if the session content is influenced by a prior request parameter. This is a secondary, less certain mapping; CAPEC-6 remains the primary pattern. Included because the content-type confusion path is a distinct mechanism from pure DOM manipulation.
Assumptions to verify
  • The affected_version field in the metadata is recorded as '0', which is treated as a placeholder meaning all versions prior to 4.1.0 are affected; no earlier version boundary is specified.
  • The CVSS PR:L assumes the attacker needs an authenticated account to create the malicious case content; if unauthenticated import paths exist, PR could be N.
  • The CVSS UI:A assumes the victim must navigate to the affected page; if the malicious content is auto-loaded in a dashboard or notification, UI could be N.
  • CAPEC-1 (Reflected XSS) is included as a secondary mapping for the content-type confusion path; the primary and more certain mapping is CAPEC-6 (DOM Based XSS).
  • The nested-JSON rejection (RecursionError guard) suggests a potential DoS vector via deeply nested JSON, but no separate DoS impact is scored because the primary vulnerability is XSS.
  • No public exploit or PoC is referenced; exploitation status is assumed to be 'none'.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 6 11 medium 6
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-10-30 11:49 UTC Fix developed Corrective change authored (5f745040d79d2f08d4a06512a91313c5a4eb346b): fix: [security] Timeline and search rendering preserve text https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b.patch
  2. 2026-11-20 10:49 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133667",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:26:26.477604Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:26:39.347Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/analyzer/misp_modules.py",
            "app/static/js/case/case_history.js",
            "app/static/js/case/timeline_view.js",
            "app/static/js/components/smart-render.js"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "flowintel contains multiple related flaws that allow an authenticated attacker to inject arbitrary HTML or script into pages viewed by other users.\n\n- Timeline rendering: the timeline library\u0027s autolinking feature rewrote already-sanitised case notes and imported MISP attribute values into generated HTML link elements. Crafted text containing quote characters could break out of the generated attribute context, altering the surrounding markup and enabling script execution in the victim\u0027s browser.\n\n- Search highlighting: the highlight injection logic used a regular expression to locate text between angle brackets and spliced \u003cmark\u003e elements into the resulting string. Because the pattern could match inside HTML attribute values, the rewrite could corrupt or inject content into attributes of already-sanitised markup, again leading to DOM-based cross-site scripting.\n\n- Content-Type confusion: three analyser selection endpoints returned session-stored JSON strings with a text/html content type. A browser receiving such a response will parse and render the body as HTML, so any HTML or script embedded in the stored value is executed in the victim\u0027s context.\n\nPreconditions: the attacker must be an authenticated user with the ability to create or import case content (notes, MISP attributes) containing specially crafted quote or angle-bracket sequences. The victim must subsequently open the affected case timeline, run a search, or request one of the selection endpoints.\n\nImpact: successful exploitation yields DOM-based cross-site scripting in the victim\u0027s browser session, potentially allowing disclosure of case data, session manipulation, or execution of arbitrary actions under the victim\u0027s identity.\n\nAffected versions: flowintel prior to 4.1.0."
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-6",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-6 DOM Based XSS"
            }
          ]
        },
        {
          "capecId": "CAPEC-1",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-1 XSS - Reflected"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 4.8,
            "baseSeverity": "MEDIUM",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "LOW",
            "userInteraction": "ACTIVE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:A/VC:L/VI:L/VA:N/SC:N/SI:L/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "LOW",
            "vulnIntegrityImpact": "LOW",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T10:50:21Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-79",
              "description": "CWE-79 Improper Neutralization of Input in Web Page (\u0027Cross-site Scripting\u0027)",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-799",
              "description": "CWE-799 Improper Control of Resources Identifiers",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T10:58:27.052Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The vulnerability is remediated by eliminating the three injection paths: (1) autolinking is disabled on every timeline event so the library no longer rewrites sanitised HTML; (2) search and auto-highlight marks are now created by walking the DOM text nodes and setting textContent, which cannot alter existing attributes or split character entities; (3) the three selection endpoints now decode the stored JSON once, respond with application/json and X-Content-Type-Options: nosniff, and reject malformed or excessively nested payloads with a 400 JSON error, preventing the browser from interpreting the body as HTML."
        }
      ],
      "title": "flowintel DOM-based XSS via Timeline Autolinking, Search Highlighting, and Content-Type Confusion",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected_version field in the metadata is recorded as \u00270\u0027, which is treated as a placeholder meaning all versions prior to 4.1.0 are affected; no earlier version boundary is specified.",
                  "The CVSS PR:L assumes the attacker needs an authenticated account to create the malicious case content; if unauthenticated import paths exist, PR could be N.",
                  "The CVSS UI:A assumes the victim must navigate to the affected page; if the malicious content is auto-loaded in a dashboard or notification, UI could be N.",
                  "CAPEC-1 (Reflected XSS) is included as a secondary mapping for the content-type confusion path; the primary and more certain mapping is CAPEC-6 (DOM Based XSS).",
                  "The nested-JSON rejection (RecursionError guard) suggests a potential DoS vector via deeply nested JSON, but no separate DoS impact is scored because the primary vulnerability is XSS.",
                  "No public exploit or PoC is referenced; exploitation status is assumed to be \u0027none\u0027."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-6",
                    "rationale": "The primary exploitation path is client-side: attacker-controlled text (case notes, MISP values, search terms) is processed by JavaScript in the browser (TimelineJS autolinking, regex-based highlight injection) and the resulting DOM manipulation injects executable markup. The content-type confusion endpoint is a secondary vector but the dominant pattern is DOM-based script injection. CAPEC-6 is the closest standard attack pattern; the mapping is confident because the patch explicitly addresses client-side HTML rewriting of user-supplied content."
                  },
                  {
                    "capecId": "CAPEC-1",
                    "rationale": "The three selection endpoints that returned session JSON as text/html could be leveraged as a reflected XSS vector if the session content is influenced by a prior request parameter. This is a secondary, less certain mapping; CAPEC-6 remains the primary pattern. Included because the content-type confusion path is a distinct mechanism from pure DOM manipulation."
                  }
                ],
                "commit": "5f745040d79d2f08d4a06512a91313c5a4eb346b",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N \u2013 the application is a web service reachable over the network. AC:L \u2013 crafting quote-bearing text in case notes or MISP attributes is straightforward; no race or special timing is needed. AT:N \u2013 no active attack or prior compromise is required beyond normal authenticated use. PR:L \u2013 the attacker needs an authenticated account to create or import the malicious content (endpoints are login_required). UI:A \u2013 the victim must actively open the case timeline, run a search, or request a selection endpoint for the injected markup to execute. VC:L \u2013 XSS can expose case data, session tokens, or other sensitive information visible in the browser. VI:L \u2013 the attacker can modify displayed content or trigger actions in the victim\u0027s session. VA:N \u2013 no evidence of denial-of-service impact beyond the nested-JSON guard. SC:N \u2013 no cross-component or cross-origin impact is demonstrated. SI:L \u2013 other authenticated users viewing the same case could be affected by the stored malicious content. SA:N \u2013 no elevation of security authority is indicated.",
                "fixSummary": "The vulnerability is remediated by eliminating the three injection paths: (1) autolinking is disabled on every timeline event so the library no longer rewrites sanitised HTML; (2) search and auto-highlight marks are now created by walking the DOM text nodes and setting textContent, which cannot alter existing attributes or split character entities; (3) the three selection endpoints now decode the stored JSON once, respond with application/json and X-Content-Type-Options: nosniff, and reject malformed or excessively nested payloads with a 400 JSON error, preventing the browser from interpreting the body as HTML.",
                "generatedAt": "2026-11-20T10:50:21.529700Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 6,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 6
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "5352fc62c6a6482c29f863cbf7a788fdb94c38d87b96d115294e35c697eeefc6",
                "patchSummary": "In case_history.js and timeline_view.js, an autolink: false property is added to each TimelineJS event object to suppress linkification. In smart-render.js, the previous regex-based string replacement (html.replace(/\u003e([^\u003c]*)\u003c/g, ...)) is replaced with a DOM TreeWalker that iterates text nodes, builds \u003cmark\u003e elements via document.createElement and textContent, and splices them into a DocumentFragment, preserving all existing attributes and entities. A new count_html_matches helper counts matches across text nodes. The plain-text fallback branches that performed raw regex replacement on escaped strings are removed. In misp_modules.py, a new _selected_json_response helper decodes the session value with json.loads, wraps it in jsonify, sets X-Content-Type-Options: nosniff, and returns a 400 JSON error on TypeError, ValueError, or RecursionError. The three get_*_selected endpoints are updated to call this helper instead of returning the raw session string.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "5f745040d79d2f08d4a06512a91313c5a4eb346b",
                    "date": "Fri, 30 Oct 2026 12:49:38 +0100",
                    "patchSha256": "5352fc62c6a6482c29f863cbf7a788fdb94c38d87b96d115294e35c697eeefc6",
                    "retrievedAt": "2026-11-20T10:49:45.815552Z",
                    "source": "https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b.patch",
                    "subject": "fix: [security] Timeline and search rendering preserve text"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T10:50:21Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Timeline and search rendering preserve text",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-79",
                    "rationale": "All three sub-issues (autolinking rewriting sanitised HTML, regex-based highlight injection corrupting attributes, and serving JSON as text/html) result in attacker-controlled content being interpreted as HTML/JavaScript in the victim\u0027s browser, which is the defining characteristic of DOM-based XSS."
                  },
                  {
                    "cweId": "CWE-799",
                    "rationale": "The content-type confusion (serving application/json data as text/html) is a secondary weakness that amplifies the XSS by causing the browser to parse the response body as HTML. This is a narrower, supporting CWE."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (5f745040d79d2f08d4a06512a91313c5a4eb346b): fix: [security] Timeline and search rendering preserve text",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b.patch"
                    ],
                    "timestamp": "2026-10-30T11:49:38Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/5f745040d79d2f08d4a06512a91313c5a4eb346b.patch"
                    ],
                    "timestamp": "2026-11-20T10:49:45.815552Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20406"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133667",
    "datePublished": "2026-11-20T10:58:27.052Z",
    "dateReserved": "2026-11-20T10:58:25.195Z",
    "dateUpdated": "2026-11-20T18:26:39.347Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133665 (GCVE-0-2026-133665)

Vulnerability from cvelistv5 – Published: 2026-11-20 10:58 – Updated: 2026-11-20 18:26
VLAI
Title
flowintel Information Disclosure of Webhook Secrets and Logs to Non-Admin Users
Summary
flowintel contains an information disclosure vulnerability that allows authenticated non-administrative users to access sensitive configuration data and logs intended for administrators only. The notification page and alert API responses exposed sensitive fields, including webhook URLs, signing secrets, IMAP credentials, and delivery logs, in the initial data or JSON responses. Although the user interface hid the configuration form from non-admin users, the underlying data remained accessible via the page source or API endpoints. This exposure included historical webhook URLs that may contain active bearer tokens. An attacker with a valid non-admin account could retrieve these credentials and logs, potentially leading to unauthorized access to notification services, impersonation of the application in webhook deliveries, or further lateral movement within the infrastructure. Affected versions: flowintel < 4.1.0
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 10:47 UTC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:25 UTC
CWE
  • CWE-200 - Exposure of Sensitive Information to an Unauthorized Actor
  • CWE-862 - Missing Authorization
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 10:47
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66.patch ee51a7b891c2…
Confidence
high
Commit Subject Patch SHA-256
1445157891ad fix: [security] Alert secrets and webhook logs stay ee51a7b891c2…
Fix summary

The vulnerability is remediated by enforcing role-based access control on data serialization and template rendering. Non-admin users are now restricted to receiving only public enable flags and do not have access to webhook URLs, secrets, or delivery logs. The shared alert serializer explicitly removes sensitive fields for non-admin users, ensuring that both the notification page and the alerts API preserve administrative diagnostics while protecting sensitive data from unauthorized access.

Patch summary

The patch modifies app/alerts/alerts_core.py and app/notification/notification.py. In alerts_core.py, the alerts_to_json function now checks if the user is an admin and removes the webhook_url field from the JSON response for non-admin users. In notification.py, the index route now checks the current user's admin status. For non-admins, it passes a restricted configuration object containing only public enable flags (CASE_CREATE_ALERT_ENABLED, WEBHOOK_ENABLED) instead of the full ConfigModule, and it passes an empty list for alert_logs instead of the actual log data. This ensures sensitive configuration and logs are not embedded in the template context for unauthorized users.

CVSS rationale

Attack Vector is Network (AV:N) as the application is web-based. Attack Complexity is Low (AC:L) as it requires only standard HTTP requests. Attack Requirements are None (AT:N). Privileges Required are Low (PR:L) as any authenticated non-admin user can exploit this. User Interaction is None (UI:N). Vulnerable Component Impact is None (VC:N) as the server itself is not compromised. Vulnerable Component Integrity is High (VI:H) because sensitive secrets and logs are exposed, which can lead to further compromise. Vulnerable Component Availability is None (VA:N). Subsequent Component Impact is None (SC:N, SI:N, SA:N) as the direct impact is limited to information disclosure within the vulnerable component, though downstream impacts are possible but not directly measured by the base vector.

Weakness rationale
  • CWE-200 The application exposed sensitive configuration data (webhook URLs, secrets, logs) to authenticated users who did not have the necessary privileges to view them.
  • CWE-862 The application failed to enforce authorization checks when serializing alert data and rendering the notification page, allowing non-admin users to access admin-only data.
Attack pattern rationale
  • CAPEC-126 The attacker leveraged their authenticated non-admin role to access data (webhook secrets, logs) that was not properly restricted to the administrator role, despite the UI hiding the configuration form.
Assumptions to verify
  • The affected version is assumed to be all versions prior to 4.1.0 based on the fixed_version metadata.
  • The CAPEC mapping is based on the general pattern of accessing restricted functionality, as the specific attack vector is a standard HTTP request to an existing endpoint.
  • The CVSS vector assumes that the exposure of secrets leads to a high impact on integrity of the vulnerable component, as these secrets can be used to impersonate the application or access other services.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 10 11 high 3
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-10-30 13:18 UTC Fix developed Corrective change authored (1445157891add7c04a63e0014cd96bf9802cac66): fix: [security] Alert secrets and webhook logs stay https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66.patch
  2. 2026-11-20 10:47 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133665",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:25:43.144657Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:26:03.887Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/alerts/alerts_core.py",
            "app/notification/notification.py"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "flowintel contains an information disclosure vulnerability that allows authenticated non-administrative users to access sensitive configuration data and logs intended for administrators only.\n\nThe notification page and alert API responses exposed sensitive fields, including webhook URLs, signing secrets, IMAP credentials, and delivery logs, in the initial data or JSON responses. Although the user interface hid the configuration form from non-admin users, the underlying data remained accessible via the page source or API endpoints. This exposure included historical webhook URLs that may contain active bearer tokens.\n\nAn attacker with a valid non-admin account could retrieve these credentials and logs, potentially leading to unauthorized access to notification services, impersonation of the application in webhook deliveries, or further lateral movement within the infrastructure.\n\nAffected versions: flowintel \u003c 4.1.0"
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-126",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-126 Accessing Functionality Not Properly Restricted by User Role"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 7.1,
            "baseSeverity": "HIGH",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:H/VA:N/SC:N/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "HIGH",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T10:47:41Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-200",
              "description": "CWE-200 Exposure of Sensitive Information to an Unauthorized Actor",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-862",
              "description": "CWE-862 Missing Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T10:58:14.508Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The vulnerability is remediated by enforcing role-based access control on data serialization and template rendering. Non-admin users are now restricted to receiving only public enable flags and do not have access to webhook URLs, secrets, or delivery logs. The shared alert serializer explicitly removes sensitive fields for non-admin users, ensuring that both the notification page and the alerts API preserve administrative diagnostics while protecting sensitive data from unauthorized access."
        }
      ],
      "title": "flowintel Information Disclosure of Webhook Secrets and Logs to Non-Admin Users",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected version is assumed to be all versions prior to 4.1.0 based on the fixed_version metadata.",
                  "The CAPEC mapping is based on the general pattern of accessing restricted functionality, as the specific attack vector is a standard HTTP request to an existing endpoint.",
                  "The CVSS vector assumes that the exposure of secrets leads to a high impact on integrity of the vulnerable component, as these secrets can be used to impersonate the application or access other services."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-126",
                    "rationale": "The attacker leveraged their authenticated non-admin role to access data (webhook secrets, logs) that was not properly restricted to the administrator role, despite the UI hiding the configuration form."
                  }
                ],
                "commit": "1445157891add7c04a63e0014cd96bf9802cac66",
                "confidence": "high",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "Attack Vector is Network (AV:N) as the application is web-based. Attack Complexity is Low (AC:L) as it requires only standard HTTP requests. Attack Requirements are None (AT:N). Privileges Required are Low (PR:L) as any authenticated non-admin user can exploit this. User Interaction is None (UI:N). Vulnerable Component Impact is None (VC:N) as the server itself is not compromised. Vulnerable Component Integrity is High (VI:H) because sensitive secrets and logs are exposed, which can lead to further compromise. Vulnerable Component Availability is None (VA:N). Subsequent Component Impact is None (SC:N, SI:N, SA:N) as the direct impact is limited to information disclosure within the vulnerable component, though downstream impacts are possible but not directly measured by the base vector.",
                "fixSummary": "The vulnerability is remediated by enforcing role-based access control on data serialization and template rendering. Non-admin users are now restricted to receiving only public enable flags and do not have access to webhook URLs, secrets, or delivery logs. The shared alert serializer explicitly removes sensitive fields for non-admin users, ensuring that both the notification page and the alerts API preserve administrative diagnostics while protecting sensitive data from unauthorized access.",
                "generatedAt": "2026-11-20T10:47:41.692593Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 3,
                      "confidence": "high",
                      "model": "qwen3.8:27b",
                      "score": 10
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "ee51a7b891c2ec45c5c6526b8a6ac91b2026442d6662c5e5888f49a807834861",
                "patchSummary": "The patch modifies app/alerts/alerts_core.py and app/notification/notification.py. In alerts_core.py, the alerts_to_json function now checks if the user is an admin and removes the webhook_url field from the JSON response for non-admin users. In notification.py, the index route now checks the current user\u0027s admin status. For non-admins, it passes a restricted configuration object containing only public enable flags (CASE_CREATE_ALERT_ENABLED, WEBHOOK_ENABLED) instead of the full ConfigModule, and it passes an empty list for alert_logs instead of the actual log data. This ensures sensitive configuration and logs are not embedded in the template context for unauthorized users.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "1445157891add7c04a63e0014cd96bf9802cac66",
                    "date": "Fri, 30 Oct 2026 14:18:20 +0100",
                    "patchSha256": "ee51a7b891c2ec45c5c6526b8a6ac91b2026442d6662c5e5888f49a807834861",
                    "retrievedAt": "2026-11-20T10:47:11.460773Z",
                    "source": "https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66.patch",
                    "subject": "fix: [security] Alert secrets and webhook logs stay"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T10:47:41Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Alert secrets and webhook logs stay",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-200",
                    "rationale": "The application exposed sensitive configuration data (webhook URLs, secrets, logs) to authenticated users who did not have the necessary privileges to view them."
                  },
                  {
                    "cweId": "CWE-862",
                    "rationale": "The application failed to enforce authorization checks when serializing alert data and rendering the notification page, allowing non-admin users to access admin-only data."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (1445157891add7c04a63e0014cd96bf9802cac66): fix: [security] Alert secrets and webhook logs stay",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66.patch"
                    ],
                    "timestamp": "2026-10-30T13:18:20Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/1445157891add7c04a63e0014cd96bf9802cac66.patch"
                    ],
                    "timestamp": "2026-11-20T10:47:11.460773Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20405"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133665",
    "datePublished": "2026-11-20T10:58:14.508Z",
    "dateReserved": "2026-11-20T10:58:12.873Z",
    "dateUpdated": "2026-11-20T18:26:03.887Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133663 (GCVE-0-2026-133663)

Vulnerability from cvelistv5 – Published: 2026-11-20 10:58 – Updated: 2026-11-20 18:25
VLAI
Title
flowintel Insecure Direct Object Reference Allows Cross-Case Data Access and Modification
Summary
flowintel contains an Insecure Direct Object Reference (IDOR) vulnerability affecting multiple endpoints that manage child records of cases and tasks. An authenticated user with access to one case or task could read or modify records belonging to other cases by supplying identifiers of those records directly in the request. Affected operations include: - Reading task notes, subtasks, and URL tools by referencing their IDs without verifying they belong to the authorized task - Retrieving task assignees for a task belonging to a different case - Editing or deleting timeline events and timeline event links belonging to another case - Editing or removing case connectors belonging to another case - Invoking connector module operations (MISP, generic connectors, worker calls) by combining an authorized case with a connector link from a different case The root cause is that the application authorized the parent resource (case or task) in the URL but then fetched or operated on the child record independently without verifying the child's ownership relationship to the authorized parent. Security impact: an authenticated user can exfiltrate sensitive threat intelligence data from cases they do not have access to, and can modify or delete timeline entries, connector configurations, and other records in unauthorized cases, leading to data integrity loss and potential disruption of incident response workflows. Affected versions: flowintel prior to 4.1.0
SSVC
Exploitation: none Automatable: yes Technical Impact: total
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 10:47 UTC
Exploitation: none Automatable: no Technical Impact: total
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:24 UTC
CWE
  • CWE-639 - Authorization Bypass Through User-Controlled Key
  • CWE-862 - Missing Authorization
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 10:47
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae.patch 5d78368cb0dd…
Confidence
high
Commit Subject Patch SHA-256
fecba79506de fix: [security] Child records stay bound to their authorised 5d78368cb0dd…
Fix summary

The fix enforces ownership validation by requiring the authorized parent case or task ID in every core method that reads or writes child records. Before returning data or performing writes, each child record's case_id or task_id is explicitly compared against the resolved parent. Mismatched records are rejected with a 404 or 403 response. Connector module context building now verifies connector-to-case ownership before preparing module data, resolving credentials, or executing sync operations. Instance visibility checks are applied consistently across both web and API routes.

Patch summary

Added case_id or task_id ownership checks in CaseCore.py methods (remove_connector, edit_connector, edit_timeline_event, edit_timeline_event_link, _build_case_connector_module_context) by passing the authorized parent ID and comparing it against the child record's foreign key. Updated case.py get_assigned_users to verify task.case_id matches the resolved case. Updated case_api.py EditConnectorsCase and RemoveConnectors to validate connector ownership and instance visibility before delegating. Updated case_misp.py timeline and connector endpoints to pass case.id to core methods. Updated task_api.py note, subtask, and URL-tool GET handlers to verify the child's task_id matches the authorized task.

CVSS rationale

AV:N - exploitable over the network via HTTP API. AC:L - the attacker only needs to know or enumerate valid child record IDs; no special race conditions or complex setup required. AT:N - no attack target manipulation needed. PR:L - requires an authenticated user with access to at least one case or task. UI:N - no user interaction required. VC:H - full read access to sensitive threat intelligence data in other cases. VI:H - ability to modify or delete timeline events, links, and connectors in other cases. VA:N - no denial-of-service impact. SC/SI/SA:N - impact is confined to the flowintel application's own data store; no external system impact is evidenced.

Weakness rationale
  • CWE-639 The vulnerability is a classic IDOR: user-supplied identifiers for child records (notes, subtasks, connectors, timeline events) are used to fetch or modify objects without verifying the object belongs to the authorized parent case or task. The authorization check on the parent does not constrain the child lookup.
  • CWE-862 Several endpoints performed the operation on a child record without any ownership or authorization check tying the record to the user's authorized case, effectively missing the authorization step for the specific object being accessed.
Attack pattern rationale
  • CAPEC-1 The attacker manipulates URL path parameters or query string identifiers (note_id, subtask_id, url_tool_id, connector instance ID, timeline event ID) to reference objects belonging to other cases or tasks. The application did not validate that the referenced object belonged to the authorized parent, allowing cross-tenant data access. This is the closest CAPEC pattern; the attack does not involve protocol-level tampering but rather parameter substitution in REST-style URLs, which falls under the broader 'tamper with URLs' category.
Assumptions to verify
  • The affected version range is inferred from the explicit context (affected_version: 0, fixed_version: 4.1.0); no earlier version boundary is confirmed.
  • PR:L assumes the attacker needs at least one valid case or task membership; the patch does not clarify whether anonymous access is possible.
  • CAPEC-1 (Tamper with URLs) is the closest available pattern; the actual attack is parameter substitution in REST endpoints rather than classic URL tampering, but no more specific CAPEC exists for IDOR via REST identifiers.
  • VC:H and VI:H assume the data in other cases is sensitive and that modification/deletion constitutes full integrity loss; the patch does not specify data classification.
  • The patch does not indicate whether the vulnerability was disclosed publicly or under embargo; exploitation status is assumed to be 'none' based on absence of evidence.
  • Instance visibility enforcement (can_interact_with_case_connector) was added to the API route in case_api.py but was already present in case_misp.py; the patch ensures consistency rather than introducing a new check in all locations.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 7 11 high 6
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-10-30 13:58 UTC Fix developed Corrective change authored (fecba79506dea2886118fe5cd686d4edfdef60ae): fix: [security] Child records stay bound to their authorised https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae.patch
  2. 2026-11-20 10:46 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133663",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "total"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:24:45.904432Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:25:02.848Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/case/CaseCore.py",
            "app/case/case.py",
            "app/case/case_api.py",
            "app/case/case_misp.py",
            "app/case/task_api.py"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "flowintel contains an Insecure Direct Object Reference (IDOR) vulnerability affecting multiple endpoints that manage child records of cases and tasks. An authenticated user with access to one case or task could read or modify records belonging to other cases by supplying identifiers of those records directly in the request.\n\nAffected operations include:\n- Reading task notes, subtasks, and URL tools by referencing their IDs without verifying they belong to the authorized task\n- Retrieving task assignees for a task belonging to a different case\n- Editing or deleting timeline events and timeline event links belonging to another case\n- Editing or removing case connectors belonging to another case\n- Invoking connector module operations (MISP, generic connectors, worker calls) by combining an authorized case with a connector link from a different case\n\nThe root cause is that the application authorized the parent resource (case or task) in the URL but then fetched or operated on the child record independently without verifying the child\u0027s ownership relationship to the authorized parent.\n\nSecurity impact: an authenticated user can exfiltrate sensitive threat intelligence data from cases they do not have access to, and can modify or delete timeline entries, connector configurations, and other records in unauthorized cases, leading to data integrity loss and potential disruption of incident response workflows.\n\nAffected versions: flowintel prior to 4.1.0"
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-1",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-1 Tamper with URLs"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 8.6,
            "baseSeverity": "HIGH",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:N/SC:N/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "HIGH",
            "vulnIntegrityImpact": "HIGH",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "total"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T10:47:07Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-639",
              "description": "CWE-639 Authorization Bypass Through User-Controlled Key",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-862",
              "description": "CWE-862 Missing Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T10:58:03.710Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The fix enforces ownership validation by requiring the authorized parent case or task ID in every core method that reads or writes child records. Before returning data or performing writes, each child record\u0027s case_id or task_id is explicitly compared against the resolved parent. Mismatched records are rejected with a 404 or 403 response. Connector module context building now verifies connector-to-case ownership before preparing module data, resolving credentials, or executing sync operations. Instance visibility checks are applied consistently across both web and API routes."
        }
      ],
      "title": "flowintel Insecure Direct Object Reference Allows Cross-Case Data Access and Modification",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected version range is inferred from the explicit context (affected_version: 0, fixed_version: 4.1.0); no earlier version boundary is confirmed.",
                  "PR:L assumes the attacker needs at least one valid case or task membership; the patch does not clarify whether anonymous access is possible.",
                  "CAPEC-1 (Tamper with URLs) is the closest available pattern; the actual attack is parameter substitution in REST endpoints rather than classic URL tampering, but no more specific CAPEC exists for IDOR via REST identifiers.",
                  "VC:H and VI:H assume the data in other cases is sensitive and that modification/deletion constitutes full integrity loss; the patch does not specify data classification.",
                  "The patch does not indicate whether the vulnerability was disclosed publicly or under embargo; exploitation status is assumed to be \u0027none\u0027 based on absence of evidence.",
                  "Instance visibility enforcement (can_interact_with_case_connector) was added to the API route in case_api.py but was already present in case_misp.py; the patch ensures consistency rather than introducing a new check in all locations."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-1",
                    "rationale": "The attacker manipulates URL path parameters or query string identifiers (note_id, subtask_id, url_tool_id, connector instance ID, timeline event ID) to reference objects belonging to other cases or tasks. The application did not validate that the referenced object belonged to the authorized parent, allowing cross-tenant data access. This is the closest CAPEC pattern; the attack does not involve protocol-level tampering but rather parameter substitution in REST-style URLs, which falls under the broader \u0027tamper with URLs\u0027 category."
                  }
                ],
                "commit": "fecba79506dea2886118fe5cd686d4edfdef60ae",
                "confidence": "high",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N - exploitable over the network via HTTP API. AC:L - the attacker only needs to know or enumerate valid child record IDs; no special race conditions or complex setup required. AT:N - no attack target manipulation needed. PR:L - requires an authenticated user with access to at least one case or task. UI:N - no user interaction required. VC:H - full read access to sensitive threat intelligence data in other cases. VI:H - ability to modify or delete timeline events, links, and connectors in other cases. VA:N - no denial-of-service impact. SC/SI/SA:N - impact is confined to the flowintel application\u0027s own data store; no external system impact is evidenced.",
                "fixSummary": "The fix enforces ownership validation by requiring the authorized parent case or task ID in every core method that reads or writes child records. Before returning data or performing writes, each child record\u0027s case_id or task_id is explicitly compared against the resolved parent. Mismatched records are rejected with a 404 or 403 response. Connector module context building now verifies connector-to-case ownership before preparing module data, resolving credentials, or executing sync operations. Instance visibility checks are applied consistently across both web and API routes.",
                "generatedAt": "2026-11-20T10:47:07.135346Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 6,
                      "confidence": "high",
                      "model": "qwen3.8:27b",
                      "score": 7
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "5d78368cb0ddd8fd9dab90b7c709b83b5aff34787ec2c25e7ba984947f4454c9",
                "patchSummary": "Added case_id or task_id ownership checks in CaseCore.py methods (remove_connector, edit_connector, edit_timeline_event, edit_timeline_event_link, _build_case_connector_module_context) by passing the authorized parent ID and comparing it against the child record\u0027s foreign key. Updated case.py get_assigned_users to verify task.case_id matches the resolved case. Updated case_api.py EditConnectorsCase and RemoveConnectors to validate connector ownership and instance visibility before delegating. Updated case_misp.py timeline and connector endpoints to pass case.id to core methods. Updated task_api.py note, subtask, and URL-tool GET handlers to verify the child\u0027s task_id matches the authorized task.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "fecba79506dea2886118fe5cd686d4edfdef60ae",
                    "date": "Fri, 30 Oct 2026 14:58:13 +0100",
                    "patchSha256": "5d78368cb0ddd8fd9dab90b7c709b83b5aff34787ec2c25e7ba984947f4454c9",
                    "retrievedAt": "2026-11-20T10:46:36.448315Z",
                    "source": "https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae.patch",
                    "subject": "fix: [security] Child records stay bound to their authorised"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "total"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T10:47:07Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Child records stay bound to their authorised",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-639",
                    "rationale": "The vulnerability is a classic IDOR: user-supplied identifiers for child records (notes, subtasks, connectors, timeline events) are used to fetch or modify objects without verifying the object belongs to the authorized parent case or task. The authorization check on the parent does not constrain the child lookup."
                  },
                  {
                    "cweId": "CWE-862",
                    "rationale": "Several endpoints performed the operation on a child record without any ownership or authorization check tying the record to the user\u0027s authorized case, effectively missing the authorization step for the specific object being accessed."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (fecba79506dea2886118fe5cd686d4edfdef60ae): fix: [security] Child records stay bound to their authorised",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae.patch"
                    ],
                    "timestamp": "2026-10-30T13:58:13Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/fecba79506dea2886118fe5cd686d4edfdef60ae.patch"
                    ],
                    "timestamp": "2026-11-20T10:46:36.448315Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20404"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133663",
    "datePublished": "2026-11-20T10:58:03.710Z",
    "dateReserved": "2026-11-20T10:58:01.964Z",
    "dateUpdated": "2026-11-20T18:25:02.848Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133651 (GCVE-0-2026-133651)

Vulnerability from cvelistv5 – Published: 2026-11-20 10:24 – Updated: 2026-11-20 18:23
VLAI
Title
FlowIntel CSRF via GET-accepted destructive web routes bypassing Flask-WTF protection
Summary
FlowIntel contained a Cross-Site Request Forgery (CSRF) vulnerability affecting fourteen web routes that performed destructive or state-changing operations. The affected routes accepted HTTP GET requests in addition to (or instead of) POST, which bypassed the Flask-WTF CSRF token validation that is only enforced on non-GET requests. The vulnerable operations included: - Case deletion and case merging - Organisation removal from a case - Task deletion (case and template) - Case template and note template deletion - Connector, connector instance, and icon deletion - Custom tag deletion and status change - Taxonomy and galaxy status changes An attacker who could induce an authenticated user to follow a crafted link or load a page containing a cross-site navigation (e.g., an image tag or redirect) could trigger any of these destructive actions using the victim's session cookie, without the victim's explicit consent or awareness. No special privileges beyond a valid authenticated session with the appropriate role (editor or admin) were required on the victim's account. Security impact: unauthorized deletion or modification of cases, tasks, templates, connectors, icons, custom tags, taxonomies, and galaxies belonging to the victim or the organisation, resulting in loss of data integrity. Affected versions: FlowIntel prior to 4.1.0
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 10:16 UTC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:23 UTC
CWE
  • CWE-352 - Cross-Site Request Forgery (CSRF)
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 10:16
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc.patch 8d58806814d8…
Confidence
high
Commit Subject Patch SHA-256
c29a27c4973f fix: [security] Require CSRF-protected POST for destructive 8d58806814d8…
Fix summary

The vulnerability is remediated by restricting all fourteen affected web routes to accept only POST requests, ensuring that Flask-WTF's CSRF token validation is enforced. All fifteen client-side callers (JavaScript fetch calls and HTML template scripts) were updated to send POST requests with the page's CSRF token in the X-CSRFToken header, preserving legitimate user interactions while blocking cross-site forged requests.

Patch summary

Backend: Changed the methods parameter on fourteen Flask route decorators from methods=['GET'] or methods=['GET','POST'] to methods=['POST'] across app/admin/admin.py, app/case/case.py, app/case/task.py, app/connectors/connectors.py, app/custom_tags/custom_tags.py, app/templating/templating.py, and app/tools/tools.py. Frontend: Updated fifteen fetch() calls in JavaScript files and HTML templates to include method: 'POST' and headers: {'X-CSRFToken': $('#csrf_token').val()}. No changes to permission decorators, query parameters, or API-key-authenticated routes.

CVSS rationale

AV:N - exploit is over the network via a web browser. AC:L - crafting a link or page that triggers a GET request is trivial. AT:N - no attack target manipulation beyond the victim's browser. PR:N - the attacker requires no account; the victim's existing session is used. UI:A - the victim must actively navigate to or click a crafted link (commit message references 'cross-site navigation'). VC:N, VI:N, VA:N - no direct impact on the application's own confidentiality, integrity, or availability. SC:N, SI:H, SA:N - the impact is on the integrity of the victim's data (deletion/modification of cases, tasks, templates, connectors, tags, taxonomies, galaxies) with no cross-system impact.

Weakness rationale
  • CWE-352 The routes accepted GET requests, which are not subject to Flask-WTF's CSRF token check. This allowed a cross-site navigation to invoke destructive state-changing operations using the victim's authenticated session, a textbook CSRF scenario.
Attack pattern rationale
  • CAPEC-124 The attack pattern involves an attacker crafting a request (in this case a GET navigation) that is sent to the target application using the victim's authenticated session cookie, causing the application to perform an unintended state-changing action. This is the closest and most direct CAPEC match for a GET-based CSRF bypass of token validation.
Assumptions to verify
  • The affected_version '0' from explicit context is treated as a placeholder indicating all versions prior to 4.1.0; the exact earliest affected version is not specified in the patch.
  • UI:A is selected based on the commit message referencing 'cross-site navigation'; if an attacker could use a passive vector (e.g., an auto-loading image tag), UI could be P, but the evidence specifically describes navigation.
  • PR:N assumes the attacker does not need their own account; the victim's authenticated session is sufficient. If the application requires the victim to have editor or admin role, this is a precondition on the victim, not a privilege requirement for the attacker.
  • CAPEC-124 is the closest standard mapping; no more specific CAPEC exists for GET-based CSRF bypass of framework token validation.
  • The patch does not indicate whether the CSRF token was previously absent from the page or only not enforced on GET; the fix implies the token was available in the DOM but the GET method bypassed validation.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 8 11 high 5
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-02 14:03 UTC Fix developed Corrective change authored (c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc): fix: [security] Require CSRF-protected POST for destructive https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc.patch
  2. 2026-11-20 10:15 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133651",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:23:00.226813Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:23:19.024Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/admin/admin.py",
            "app/case/case.py",
            "app/case/task.py",
            "app/connectors/connectors.py",
            "app/custom_tags/custom_tags.py",
            "app/templating/templating.py",
            "app/tools/tools.py"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "FlowIntel contained a Cross-Site Request Forgery (CSRF) vulnerability affecting fourteen web routes that performed destructive or state-changing operations. The affected routes accepted HTTP GET requests in addition to (or instead of) POST, which bypassed the Flask-WTF CSRF token validation that is only enforced on non-GET requests.\n\nThe vulnerable operations included:\n- Case deletion and case merging\n- Organisation removal from a case\n- Task deletion (case and template)\n- Case template and note template deletion\n- Connector, connector instance, and icon deletion\n- Custom tag deletion and status change\n- Taxonomy and galaxy status changes\n\nAn attacker who could induce an authenticated user to follow a crafted link or load a page containing a cross-site navigation (e.g., an image tag or redirect) could trigger any of these destructive actions using the victim\u0027s session cookie, without the victim\u0027s explicit consent or awareness. No special privileges beyond a valid authenticated session with the appropriate role (editor or admin) were required on the victim\u0027s account.\n\nSecurity impact: unauthorized deletion or modification of cases, tasks, templates, connectors, icons, custom tags, taxonomies, and galaxies belonging to the victim or the organisation, resulting in loss of data integrity.\n\nAffected versions: FlowIntel prior to 4.1.0"
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-124",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-124 Cross-Site Request Forgery"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 6.1,
            "baseSeverity": "MEDIUM",
            "privilegesRequired": "NONE",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "HIGH",
            "userInteraction": "ACTIVE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:A/VC:N/VI:N/VA:N/SC:N/SI:H/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "NONE",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T10:16:24Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-352",
              "description": "CWE-352 Cross-Site Request Forgery (CSRF)",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T10:24:17.885Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The vulnerability is remediated by restricting all fourteen affected web routes to accept only POST requests, ensuring that Flask-WTF\u0027s CSRF token validation is enforced. All fifteen client-side callers (JavaScript fetch calls and HTML template scripts) were updated to send POST requests with the page\u0027s CSRF token in the X-CSRFToken header, preserving legitimate user interactions while blocking cross-site forged requests."
        }
      ],
      "title": "FlowIntel CSRF via GET-accepted destructive web routes bypassing Flask-WTF protection",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected_version \u00270\u0027 from explicit context is treated as a placeholder indicating all versions prior to 4.1.0; the exact earliest affected version is not specified in the patch.",
                  "UI:A is selected based on the commit message referencing \u0027cross-site navigation\u0027; if an attacker could use a passive vector (e.g., an auto-loading image tag), UI could be P, but the evidence specifically describes navigation.",
                  "PR:N assumes the attacker does not need their own account; the victim\u0027s authenticated session is sufficient. If the application requires the victim to have editor or admin role, this is a precondition on the victim, not a privilege requirement for the attacker.",
                  "CAPEC-124 is the closest standard mapping; no more specific CAPEC exists for GET-based CSRF bypass of framework token validation.",
                  "The patch does not indicate whether the CSRF token was previously absent from the page or only not enforced on GET; the fix implies the token was available in the DOM but the GET method bypassed validation."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-124",
                    "rationale": "The attack pattern involves an attacker crafting a request (in this case a GET navigation) that is sent to the target application using the victim\u0027s authenticated session cookie, causing the application to perform an unintended state-changing action. This is the closest and most direct CAPEC match for a GET-based CSRF bypass of token validation."
                  }
                ],
                "commit": "c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc",
                "confidence": "high",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N - exploit is over the network via a web browser. AC:L - crafting a link or page that triggers a GET request is trivial. AT:N - no attack target manipulation beyond the victim\u0027s browser. PR:N - the attacker requires no account; the victim\u0027s existing session is used. UI:A - the victim must actively navigate to or click a crafted link (commit message references \u0027cross-site navigation\u0027). VC:N, VI:N, VA:N - no direct impact on the application\u0027s own confidentiality, integrity, or availability. SC:N, SI:H, SA:N - the impact is on the integrity of the victim\u0027s data (deletion/modification of cases, tasks, templates, connectors, tags, taxonomies, galaxies) with no cross-system impact.",
                "fixSummary": "The vulnerability is remediated by restricting all fourteen affected web routes to accept only POST requests, ensuring that Flask-WTF\u0027s CSRF token validation is enforced. All fifteen client-side callers (JavaScript fetch calls and HTML template scripts) were updated to send POST requests with the page\u0027s CSRF token in the X-CSRFToken header, preserving legitimate user interactions while blocking cross-site forged requests.",
                "generatedAt": "2026-11-20T10:16:24.400257Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 5,
                      "confidence": "high",
                      "model": "qwen3.8:27b",
                      "score": 8
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "8d58806814d8ac1f104c7206ff3b4c2e842c8a0c94205dcf816813439ac98863",
                "patchSummary": "Backend: Changed the methods parameter on fourteen Flask route decorators from methods=[\u0027GET\u0027] or methods=[\u0027GET\u0027,\u0027POST\u0027] to methods=[\u0027POST\u0027] across app/admin/admin.py, app/case/case.py, app/case/task.py, app/connectors/connectors.py, app/custom_tags/custom_tags.py, app/templating/templating.py, and app/tools/tools.py. Frontend: Updated fifteen fetch() calls in JavaScript files and HTML templates to include method: \u0027POST\u0027 and headers: {\u0027X-CSRFToken\u0027: $(\u0027#csrf_token\u0027).val()}. No changes to permission decorators, query parameters, or API-key-authenticated routes.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc",
                    "date": "Mon, 2 Nov 2026 15:03:14 +0100",
                    "patchSha256": "8d58806814d8ac1f104c7206ff3b4c2e842c8a0c94205dcf816813439ac98863",
                    "retrievedAt": "2026-11-20T10:15:57.346441Z",
                    "source": "https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc.patch",
                    "subject": "fix: [security] Require CSRF-protected POST for destructive"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T10:16:24Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Require CSRF-protected POST for destructive",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-352",
                    "rationale": "The routes accepted GET requests, which are not subject to Flask-WTF\u0027s CSRF token check. This allowed a cross-site navigation to invoke destructive state-changing operations using the victim\u0027s authenticated session, a textbook CSRF scenario."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc): fix: [security] Require CSRF-protected POST for destructive",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc.patch"
                    ],
                    "timestamp": "2026-11-02T14:03:14Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/c29a27c4973f0e47cb235cdfbeb8743bd11a3bbc.patch"
                    ],
                    "timestamp": "2026-11-20T10:15:57.346441Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20237"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133651",
    "datePublished": "2026-11-20T10:24:17.885Z",
    "dateReserved": "2026-11-20T10:24:15.465Z",
    "dateUpdated": "2026-11-20T18:23:19.024Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133648 (GCVE-0-2026-133648)

Vulnerability from cvelistv5 – Published: 2026-11-20 10:23 – Updated: 2026-11-20 18:22
VLAI
Title
Flowintel Disables TLS Certificate Verification for MISP and Mail Connections
Summary
Flowintel disabled TLS certificate verification for all MISP API connections and used unverified SSL contexts for IMAP and SMTP STARTTLS mail connections. This allowed an on-path (man-in-the-middle) attacker positioned between Flowintel and its MISP or mail servers to impersonate the remote server and intercept sensitive data. Affected connections include: - All MISP HTTPS API calls (event retrieval, attribute search, event creation, connectivity checks, and API validation) - SMTP STARTTLS connections used for sending email notifications - IMAP TLS connections used for fetching mail archives and testing connectivity An attacker who can observe or intercept network traffic between Flowintel and the MISP or mail server could: - Capture MISP API keys and mail credentials (SMTP/IMAP passwords) - Read all threat intelligence data exchanged (events, attributes, objects, reports) - Inject or modify threat intelligence data returned to Flowintel The vulnerability required no authentication on the part of the attacker and no user interaction; it was exploitable by any party with a network position between Flowintel and its configured MISP or mail endpoints. Affected versions: Flowintel prior to 4.1.0.
SSVC
Exploitation: none Automatable: yes Technical Impact: total
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 10:15 UTC
Exploitation: none Automatable: yes Technical Impact: total
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:22 UTC
CWE
  • CWE-295 - Improper Certificate Validation
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 10:15
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f.patch 286e145e7697…
Confidence
high
Commit Subject Patch SHA-256
b8e257ab1763 fix: [security] Verify TLS certificates for MISP and mail 286e145e7697…
Fix summary

The fix introduces a shared TLS utility module that enforces certificate and hostname verification by default for all MISP and mail connections. MISP clients now receive a proper SSL verification context instead of ssl=False, and IMAP/SMTP connections use a verified ssl.SSLContext. Operators may configure a private CA bundle path or explicitly disable verification (which fails closed on malformed input). The previous urllib3 warning suppression is removed, and upgrade documentation warns that previously accepted self-signed or mismatched certificates will now cause connection failures.

Patch summary

Added new file app/utils/tls.py providing misp_ssl() and mail_ssl_context() helpers that read MISP_VERIFY_SSL and MAIL_VERIFY_SSL config values (default True) and return appropriate verification settings or SSL contexts. Replaced all ssl=False arguments in PyMISP constructor calls across 8 files with ssl=misp_ssl(). Replaced bare server.starttls() with server.starttls(context=mail_ssl_context()) and added ssl_context=mail_ssl_context() to IMAP4_SSL calls. Removed all urllib3.disable_warnings() calls. Added MISP_VERIFY_SSL and MAIL_VERIFY_SSL to the default config template. Added installation documentation for TLS configuration and upgrade notes. The _verify_setting() helper raises ValueError on invalid input, ensuring fail-closed behavior.

CVSS rationale

AV:N - The attack is performed over the network by intercepting traffic between Flowintel and MISP/mail servers. AC:L - No special race conditions or complex setup are needed; the attacker only needs a network position on the path. AT:N - Flowintel is a server application; no active user on the target is required for the connection to be made. PR:N - The attacker does not need any credentials or account on Flowintel. UI:N - No user interaction is required. VC:H - All data in transit (API keys, mail credentials, full threat intelligence content) is readable by the attacker. VI:H - The attacker can modify or inject threat intelligence data returned to Flowintel. VA:N - No denial-of-service impact is indicated. SC:H - The MISP instance's data and the mail server's data are exposed to the attacker. SI:H - Forged or modified data can be injected into the MISP instance via the compromised channel. SA:N - No security authority impact is evident.

Weakness rationale
  • CWE-295 The application explicitly disabled TLS certificate verification by passing ssl=False to all MISP client connections and by using unverified SSL contexts for IMAP/SMTP. This is a textbook case of improper certificate validation where the client accepts any server certificate without verifying the chain or hostname.
Attack pattern rationale
  • CAPEC-119 The disabled certificate verification directly enables a classic man-in-the-middle attack: an on-path attacker presents a self-signed or rogue certificate, Flowintel accepts it without validation, and the attacker reads and modifies all traffic. CAPEC-119 is the closest standard attack pattern describing interception and modification of communications between two parties. The patch commit message explicitly references an 'on-path attacker' impersonating the server, confirming this mapping.
Assumptions to verify
  • The affected version range is inferred from the explicit context (affected_version '0', fixed_version '4.1.0'); the patch itself does not enumerate all prior releases.
  • CVSS VI:H assumes the attacker can successfully inject or modify threat intelligence data in transit; if the MISP API enforces additional integrity checks server-side, the practical integrity impact may be lower.
  • CAPEC-119 (Man-in-the-Middle) is selected as the closest standard attack pattern; the patch does not reference a specific CAPEC, and the attack could also be described under CAPEC-123 (Session Hijacking) or CAPEC-107 (Web Service Cross-Site Request Forgery) in a broader sense, but CAPEC-119 most directly matches the described on-path interception scenario.
  • The patch assumes MISP connector URLs use HTTPS; if an operator configures a plain HTTP MISP URL, TLS verification is not applicable and the vulnerability does not manifest in the same way.
  • PR:N assumes the attacker does not need any Flowintel account; the vulnerability is in the outbound connection from Flowintel to MISP/mail, not in Flowintel's own authentication.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 8 11 high 5
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-02 14:48 UTC Fix developed Corrective change authored (b8e257ab17632eb37c912485cbd78ee7514b764f): fix: [security] Verify TLS certificates for MISP and mail https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f.patch
  2. 2026-11-20 10:15 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133648",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "yes"
                  },
                  {
                    "Technical Impact": "total"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:22:26.198046Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:22:41.449Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/connectors/connectors_core.py",
            "app/modules/notify_user/email.py",
            "app/modules/receive_from/receive_misp_object.py",
            "app/modules/send_to/misp_event.py",
            "app/modules/send_to/misp_event_task.py",
            "app/modules/send_to/misp_object_event.py",
            "app/tools/tools_core.py",
            "app/tools/validate_api.py"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "Flowintel disabled TLS certificate verification for all MISP API connections and used unverified SSL contexts for IMAP and SMTP STARTTLS mail connections. This allowed an on-path (man-in-the-middle) attacker positioned between Flowintel and its MISP or mail servers to impersonate the remote server and intercept sensitive data.\n\nAffected connections include:\n- All MISP HTTPS API calls (event retrieval, attribute search, event creation, connectivity checks, and API validation)\n- SMTP STARTTLS connections used for sending email notifications\n- IMAP TLS connections used for fetching mail archives and testing connectivity\n\nAn attacker who can observe or intercept network traffic between Flowintel and the MISP or mail server could:\n- Capture MISP API keys and mail credentials (SMTP/IMAP passwords)\n- Read all threat intelligence data exchanged (events, attributes, objects, reports)\n- Inject or modify threat intelligence data returned to Flowintel\n\nThe vulnerability required no authentication on the part of the attacker and no user interaction; it was exploitable by any party with a network position between Flowintel and its configured MISP or mail endpoints.\n\nAffected versions: Flowintel prior to 4.1.0."
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-119",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-119 Man-in-the-Middle"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 9.9,
            "baseSeverity": "CRITICAL",
            "privilegesRequired": "NONE",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "HIGH",
            "subIntegrityImpact": "HIGH",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:N/SC:H/SI:H/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "HIGH",
            "vulnIntegrityImpact": "HIGH",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "total"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T10:15:55Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-295",
              "description": "CWE-295 Improper Certificate Validation",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T10:23:53.537Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The fix introduces a shared TLS utility module that enforces certificate and hostname verification by default for all MISP and mail connections. MISP clients now receive a proper SSL verification context instead of ssl=False, and IMAP/SMTP connections use a verified ssl.SSLContext. Operators may configure a private CA bundle path or explicitly disable verification (which fails closed on malformed input). The previous urllib3 warning suppression is removed, and upgrade documentation warns that previously accepted self-signed or mismatched certificates will now cause connection failures."
        }
      ],
      "title": "Flowintel Disables TLS Certificate Verification for MISP and Mail Connections",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected version range is inferred from the explicit context (affected_version \u00270\u0027, fixed_version \u00274.1.0\u0027); the patch itself does not enumerate all prior releases.",
                  "CVSS VI:H assumes the attacker can successfully inject or modify threat intelligence data in transit; if the MISP API enforces additional integrity checks server-side, the practical integrity impact may be lower.",
                  "CAPEC-119 (Man-in-the-Middle) is selected as the closest standard attack pattern; the patch does not reference a specific CAPEC, and the attack could also be described under CAPEC-123 (Session Hijacking) or CAPEC-107 (Web Service Cross-Site Request Forgery) in a broader sense, but CAPEC-119 most directly matches the described on-path interception scenario.",
                  "The patch assumes MISP connector URLs use HTTPS; if an operator configures a plain HTTP MISP URL, TLS verification is not applicable and the vulnerability does not manifest in the same way.",
                  "PR:N assumes the attacker does not need any Flowintel account; the vulnerability is in the outbound connection from Flowintel to MISP/mail, not in Flowintel\u0027s own authentication."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-119",
                    "rationale": "The disabled certificate verification directly enables a classic man-in-the-middle attack: an on-path attacker presents a self-signed or rogue certificate, Flowintel accepts it without validation, and the attacker reads and modifies all traffic. CAPEC-119 is the closest standard attack pattern describing interception and modification of communications between two parties. The patch commit message explicitly references an \u0027on-path attacker\u0027 impersonating the server, confirming this mapping."
                  }
                ],
                "commit": "b8e257ab17632eb37c912485cbd78ee7514b764f",
                "confidence": "high",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N - The attack is performed over the network by intercepting traffic between Flowintel and MISP/mail servers. AC:L - No special race conditions or complex setup are needed; the attacker only needs a network position on the path. AT:N - Flowintel is a server application; no active user on the target is required for the connection to be made. PR:N - The attacker does not need any credentials or account on Flowintel. UI:N - No user interaction is required. VC:H - All data in transit (API keys, mail credentials, full threat intelligence content) is readable by the attacker. VI:H - The attacker can modify or inject threat intelligence data returned to Flowintel. VA:N - No denial-of-service impact is indicated. SC:H - The MISP instance\u0027s data and the mail server\u0027s data are exposed to the attacker. SI:H - Forged or modified data can be injected into the MISP instance via the compromised channel. SA:N - No security authority impact is evident.",
                "fixSummary": "The fix introduces a shared TLS utility module that enforces certificate and hostname verification by default for all MISP and mail connections. MISP clients now receive a proper SSL verification context instead of ssl=False, and IMAP/SMTP connections use a verified ssl.SSLContext. Operators may configure a private CA bundle path or explicitly disable verification (which fails closed on malformed input). The previous urllib3 warning suppression is removed, and upgrade documentation warns that previously accepted self-signed or mismatched certificates will now cause connection failures.",
                "generatedAt": "2026-11-20T10:15:55.134458Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 5,
                      "confidence": "high",
                      "model": "qwen3.8:27b",
                      "score": 8
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "286e145e7697fc6bd5fcba49ad4998a8b621fa9a18f2b287ba3c344f26f06338",
                "patchSummary": "Added new file app/utils/tls.py providing misp_ssl() and mail_ssl_context() helpers that read MISP_VERIFY_SSL and MAIL_VERIFY_SSL config values (default True) and return appropriate verification settings or SSL contexts. Replaced all ssl=False arguments in PyMISP constructor calls across 8 files with ssl=misp_ssl(). Replaced bare server.starttls() with server.starttls(context=mail_ssl_context()) and added ssl_context=mail_ssl_context() to IMAP4_SSL calls. Removed all urllib3.disable_warnings() calls. Added MISP_VERIFY_SSL and MAIL_VERIFY_SSL to the default config template. Added installation documentation for TLS configuration and upgrade notes. The _verify_setting() helper raises ValueError on invalid input, ensuring fail-closed behavior.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "b8e257ab17632eb37c912485cbd78ee7514b764f",
                    "date": "Mon, 2 Nov 2026 15:48:58 +0100",
                    "patchSha256": "286e145e7697fc6bd5fcba49ad4998a8b621fa9a18f2b287ba3c344f26f06338",
                    "retrievedAt": "2026-11-20T10:15:19.966655Z",
                    "source": "https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f.patch",
                    "subject": "fix: [security] Verify TLS certificates for MISP and mail"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "total"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T10:15:55Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Verify TLS certificates for MISP and mail",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-295",
                    "rationale": "The application explicitly disabled TLS certificate verification by passing ssl=False to all MISP client connections and by using unverified SSL contexts for IMAP/SMTP. This is a textbook case of improper certificate validation where the client accepts any server certificate without verifying the chain or hostname."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (b8e257ab17632eb37c912485cbd78ee7514b764f): fix: [security] Verify TLS certificates for MISP and mail",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f.patch"
                    ],
                    "timestamp": "2026-11-02T14:48:58Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/b8e257ab17632eb37c912485cbd78ee7514b764f.patch"
                    ],
                    "timestamp": "2026-11-20T10:15:19.966655Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20254"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133648",
    "datePublished": "2026-11-20T10:23:53.537Z",
    "dateReserved": "2026-11-20T10:23:51.984Z",
    "dateUpdated": "2026-11-20T18:22:41.449Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133631 (GCVE-0-2026-133631)

Vulnerability from cvelistv5 – Published: 2026-11-20 09:10 – Updated: 2026-11-20 18:18
VLAI
Title
FlowIntel Insecure Note Variable Resolution, Over-Permissive Chatbot API Access, and Credential Exposure in Process Arguments
Summary
FlowIntel contains multiple security flaws in its note variable resolution, chatbot API tooling, and AI report generation. - Note variable references (e.g. cross-case or cross-task lookups) were resolved using the viewing user's privileges rather than the destination case's audience. An authenticated user with access to a shared or imported note could embed references to private cases, causing sensitive case data to be pulled into AI report requests, exports, and MISP sends without the referenced case's audience having granted read access. - The chatbot API tool permitted the language model to invoke any FlowIntel API endpoint, including administrative routes and state-changing GET operations, without an allow-list or path validation. An attacker who can influence the model's instructions (e.g. via untrusted case content) could induce data modification or access restricted endpoints. - AI report generation passed the Ollama bearer token and full case content as command-line arguments to a child process, exposing them in the system process table to any local user. Affected versions: FlowIntel prior to 4.1.0.
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 09:04 UTC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:17 UTC
CWE
  • CWE-862 - Missing Authorization
  • CWE-284 - Improper Access Control
  • CWE-200 - Exposure of Sensitive Information to an Unauthorized Actor
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 09:04
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1.patch c2ffe93f6a63…
Confidence
medium
Commit Subject Patch SHA-256
34d151b8dceb fix: [security] Keep note references and AI tooling within c2ffe93f6a63…
Fix summary

Three remediations are applied: (1) Note variable resolution now enforces an audience-based authorization check so that cross-case and cross-task references only expand when the destination case's assigned organisations can read the referenced case; (2) the chatbot API tool is restricted to an explicit allow-list of reviewed read-only and search operations, with administrative endpoints always blocked, path validation against injection and SSRF, and write operations gated behind an explicit CHATBOT_ALLOW_WRITES configuration flag; (3) AI report requests are delivered to the child process via a private anonymous temporary file on stdin instead of argv, keeping credentials and case content out of the process listing.

Patch summary

In note_variables.py, a new _audience_can_view() function verifies that the destination case's organisations are a subset of the referenced case's organisations before resolving cross-case or cross-task references; case_notes.py passes the resolved case object instead of a raw integer ID. In chatbot_core.py, a _chatbot_api_request() validator and a _READ_API_PATHS regex allow-list are added; unsupported methods, external URLs, path traversal, admin routes, and non-allow-listed operations are refused before the MCP call. In CaseCore.py, the subprocess invocation for AI reports is changed from passing URL, payload, and headers as argv to writing them as JSON into a tempfile.TemporaryFile and feeding it via stdin. A CHATBOT_ALLOW_WRITES flag is added to the default config, and documentation is updated in the installation and user manuals.

CVSS rationale

AV:N: FlowIntel is a network-accessible web application. AC:L: An authenticated user can craft note references or interact with the chatbot without special conditions. AT:N: No specific attack target is required. PR:L: A low-privilege authenticated user with access to a shared note or the chatbot is sufficient. UI:N: No victim interaction beyond normal use is needed. VC:H: Private case data (titles, notes, MISP objects, user data) is exposed to unauthorized viewers and external AI/MISP systems. VI:L: The chatbot can invoke state-changing endpoints, allowing limited data modification. VA:N: No availability impact is evidenced. SC:L: Sensitive data is forwarded to external AI (Ollama) and MISP instances. SI:N: No subsequent integrity impact is demonstrated. SA:N: No subsequent availability impact.

Weakness rationale
  • CWE-862 Note variable resolution did not verify that the destination case's audience had read access to the referenced case, allowing an authenticated user to pull private data from cases they are not authorised to view.
  • CWE-284 The chatbot API tool lacked an allow-list or restriction on which API endpoints the model could invoke, permitting access to administrative and state-changing routes.
  • CWE-200 The Ollama bearer token and case content were passed as command-line arguments, visible in the local process table to any user on the host.
Attack pattern rationale
  • CAPEC-126 The note variable resolution relied solely on the viewer's own privileges without checking the destination case's audience, effectively an incorrectly implemented authorization scheme. An attacker crafts cross-case references in a note they can view to exfiltrate data from cases outside their access scope. This is the closest CAPEC to the missing audience-based authorization check.
  • CAPEC-12 The chatbot API tool was configured to permit unrestricted API calls without an allow-list, representing an application configuration error that exposes administrative and write endpoints to model-driven requests. The fix introduces a restrictive default configuration. This mapping is approximate; no CAPEC specifically covers LLM-driven API over-permissiveness.
Assumptions to verify
  • The affected version range is inferred from the explicit context (affected_version 0, fixed_version 4.1.0); no specific prior version boundary is confirmed in the patch.
  • PR:L assumes the attacker is an authenticated FlowIntel user with access to at least one shared or imported note and the chatbot interface; unauthenticated access is not evidenced.
  • VI:L for the chatbot assumes CHATBOT_ALLOW_WRITES is disabled by default; if enabled, the integrity impact could be higher.
  • SC:L assumes AI report data and MISP sends reach external systems; if the deployment is fully air-gapped, SC could be N.
  • CAPEC-126 is the closest available pattern for the missing audience-based authorization; no CAPEC specifically models cross-entity reference resolution authorization gaps.
  • CAPEC-12 is used for the chatbot over-permissiveness; no CAPEC specifically covers LLM-driven unrestricted API invocation.
  • The credential-in-argv issue (CWE-200) is a local information exposure and may not be exploitable in containerized or single-user deployments; its contribution to the overall CVSS is minor.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 5 11 medium 7
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-03 10:53 UTC Fix developed Corrective change authored (34d151b8dceb1fea925b353f04a9483aba0a22b1): fix: [security] Keep note references and AI tooling within https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1.patch
  2. 2026-11-20 09:03 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133631",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:17:51.923396Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:18:08.558Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/case/CaseCore.py",
            "app/case/case_notes.py",
            "app/chatbot/chatbot_core.py",
            "app/utils/note_variables.py"
          ],
          "product": "flowintel",
          "programFiles": [],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "FlowIntel contains multiple security flaws in its note variable resolution, chatbot API tooling, and AI report generation.\n\n- Note variable references (e.g. cross-case or cross-task lookups) were resolved using the viewing user\u0027s privileges rather than the destination case\u0027s audience. An authenticated user with access to a shared or imported note could embed references to private cases, causing sensitive case data to be pulled into AI report requests, exports, and MISP sends without the referenced case\u0027s audience having granted read access.\n\n- The chatbot API tool permitted the language model to invoke any FlowIntel API endpoint, including administrative routes and state-changing GET operations, without an allow-list or path validation. An attacker who can influence the model\u0027s instructions (e.g. via untrusted case content) could induce data modification or access restricted endpoints.\n\n- AI report generation passed the Ollama bearer token and full case content as command-line arguments to a child process, exposing them in the system process table to any local user.\n\nAffected versions: FlowIntel prior to 4.1.0."
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-126",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-126 Exploiting Incorrectly Implemented Authentication Schemes"
            }
          ]
        },
        {
          "capecId": "CAPEC-12",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-12 Application Configuration Error"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 7.1,
            "baseSeverity": "HIGH",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "LOW",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:L/VA:N/SC:L/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "HIGH",
            "vulnIntegrityImpact": "LOW",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T09:04:05Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-862",
              "description": "CWE-862 Missing Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-284",
              "description": "CWE-284 Improper Access Control",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-200",
              "description": "CWE-200 Exposure of Sensitive Information to an Unauthorized Actor",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T09:11:43.154Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "Three remediations are applied: (1) Note variable resolution now enforces an audience-based authorization check so that cross-case and cross-task references only expand when the destination case\u0027s assigned organisations can read the referenced case; (2) the chatbot API tool is restricted to an explicit allow-list of reviewed read-only and search operations, with administrative endpoints always blocked, path validation against injection and SSRF, and write operations gated behind an explicit CHATBOT_ALLOW_WRITES configuration flag; (3) AI report requests are delivered to the child process via a private anonymous temporary file on stdin instead of argv, keeping credentials and case content out of the process listing."
        }
      ],
      "title": "FlowIntel Insecure Note Variable Resolution, Over-Permissive Chatbot API Access, and Credential Exposure in Process Arguments",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The affected version range is inferred from the explicit context (affected_version 0, fixed_version 4.1.0); no specific prior version boundary is confirmed in the patch.",
                  "PR:L assumes the attacker is an authenticated FlowIntel user with access to at least one shared or imported note and the chatbot interface; unauthenticated access is not evidenced.",
                  "VI:L for the chatbot assumes CHATBOT_ALLOW_WRITES is disabled by default; if enabled, the integrity impact could be higher.",
                  "SC:L assumes AI report data and MISP sends reach external systems; if the deployment is fully air-gapped, SC could be N.",
                  "CAPEC-126 is the closest available pattern for the missing audience-based authorization; no CAPEC specifically models cross-entity reference resolution authorization gaps.",
                  "CAPEC-12 is used for the chatbot over-permissiveness; no CAPEC specifically covers LLM-driven unrestricted API invocation.",
                  "The credential-in-argv issue (CWE-200) is a local information exposure and may not be exploitable in containerized or single-user deployments; its contribution to the overall CVSS is minor."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-126",
                    "rationale": "The note variable resolution relied solely on the viewer\u0027s own privileges without checking the destination case\u0027s audience, effectively an incorrectly implemented authorization scheme. An attacker crafts cross-case references in a note they can view to exfiltrate data from cases outside their access scope. This is the closest CAPEC to the missing audience-based authorization check."
                  },
                  {
                    "capecId": "CAPEC-12",
                    "rationale": "The chatbot API tool was configured to permit unrestricted API calls without an allow-list, representing an application configuration error that exposes administrative and write endpoints to model-driven requests. The fix introduces a restrictive default configuration. This mapping is approximate; no CAPEC specifically covers LLM-driven API over-permissiveness."
                  }
                ],
                "commit": "34d151b8dceb1fea925b353f04a9483aba0a22b1",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N: FlowIntel is a network-accessible web application. AC:L: An authenticated user can craft note references or interact with the chatbot without special conditions. AT:N: No specific attack target is required. PR:L: A low-privilege authenticated user with access to a shared note or the chatbot is sufficient. UI:N: No victim interaction beyond normal use is needed. VC:H: Private case data (titles, notes, MISP objects, user data) is exposed to unauthorized viewers and external AI/MISP systems. VI:L: The chatbot can invoke state-changing endpoints, allowing limited data modification. VA:N: No availability impact is evidenced. SC:L: Sensitive data is forwarded to external AI (Ollama) and MISP instances. SI:N: No subsequent integrity impact is demonstrated. SA:N: No subsequent availability impact.",
                "fixSummary": "Three remediations are applied: (1) Note variable resolution now enforces an audience-based authorization check so that cross-case and cross-task references only expand when the destination case\u0027s assigned organisations can read the referenced case; (2) the chatbot API tool is restricted to an explicit allow-list of reviewed read-only and search operations, with administrative endpoints always blocked, path validation against injection and SSRF, and write operations gated behind an explicit CHATBOT_ALLOW_WRITES configuration flag; (3) AI report requests are delivered to the child process via a private anonymous temporary file on stdin instead of argv, keeping credentials and case content out of the process listing.",
                "generatedAt": "2026-11-20T09:04:05.830413Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 7,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 5
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "c2ffe93f6a63b984f6d1a733093060b3def8170a7f47b9662ca137faa3a5d7e4",
                "patchSummary": "In note_variables.py, a new _audience_can_view() function verifies that the destination case\u0027s organisations are a subset of the referenced case\u0027s organisations before resolving cross-case or cross-task references; case_notes.py passes the resolved case object instead of a raw integer ID. In chatbot_core.py, a _chatbot_api_request() validator and a _READ_API_PATHS regex allow-list are added; unsupported methods, external URLs, path traversal, admin routes, and non-allow-listed operations are refused before the MCP call. In CaseCore.py, the subprocess invocation for AI reports is changed from passing URL, payload, and headers as argv to writing them as JSON into a tempfile.TemporaryFile and feeding it via stdin. A CHATBOT_ALLOW_WRITES flag is added to the default config, and documentation is updated in the installation and user manuals.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "34d151b8dceb1fea925b353f04a9483aba0a22b1",
                    "date": "Tue, 3 Nov 2026 11:53:18 +0100",
                    "patchSha256": "c2ffe93f6a63b984f6d1a733093060b3def8170a7f47b9662ca137faa3a5d7e4",
                    "retrievedAt": "2026-11-20T09:03:27.928581Z",
                    "source": "https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1.patch",
                    "subject": "fix: [security] Keep note references and AI tooling within"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T09:04:05Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Keep note references and AI tooling within",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-862",
                    "rationale": "Note variable resolution did not verify that the destination case\u0027s audience had read access to the referenced case, allowing an authenticated user to pull private data from cases they are not authorised to view."
                  },
                  {
                    "cweId": "CWE-284",
                    "rationale": "The chatbot API tool lacked an allow-list or restriction on which API endpoints the model could invoke, permitting access to administrative and state-changing routes."
                  },
                  {
                    "cweId": "CWE-200",
                    "rationale": "The Ollama bearer token and case content were passed as command-line arguments, visible in the local process table to any user on the host."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (34d151b8dceb1fea925b353f04a9483aba0a22b1): fix: [security] Keep note references and AI tooling within",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1.patch"
                    ],
                    "timestamp": "2026-11-03T10:53:18Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/34d151b8dceb1fea925b353f04a9483aba0a22b1.patch"
                    ],
                    "timestamp": "2026-11-20T09:03:27.928581Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20355"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133631",
    "datePublished": "2026-11-20T09:10:23.474Z",
    "dateReserved": "2026-11-20T09:10:21.313Z",
    "dateUpdated": "2026-11-20T18:18:08.558Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133630 (GCVE-0-2026-133630)

Vulnerability from cvelistv5 – Published: 2026-11-20 09:09 – Updated: 2026-11-20 18:17
VLAI
Title
flowintel CI/CD Workflow Exposes Checkout Credentials in Docker Build Cache
Summary
The flowintel Docker build workflow used the GitHub Actions checkout action with default credential persistence, causing the job authentication token to be written into the repository's Git configuration file. This file became part of the Docker build context and could be retained in exported build cache layers, even though Git metadata was stripped from the final runtime image. An attacker with access to the CI/CD build cache or build environment could extract the persisted token and use it to gain unauthorized access to the private source repository. Impact: - Exposure of a valid authentication token to the private repository - Potential unauthorized read access to source code and repository metadata Affected versions: flowintel < 4.1.0
SSVC
Exploitation: none Automatable: no Technical Impact: partial
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 09:03 UTC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:16 UTC
CWE
  • CWE-522 - Insufficiently Protected Credentials
  • CWE-200 - Exposure of Sensitive Information to an Unauthorized Actor
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 09:03
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05.patch 6cdbee24e7fd…
Confidence
medium
Commit Subject Patch SHA-256
6b26b3f8cf9f fix: [security] Keep checkout credentials out of Docker build 6cdbee24e7fd…
Fix summary

The checkout step is configured to not persist credentials into the workspace, and the full branch history is fetched during the authenticated checkout phase. This ensures the job token is never written to the Git configuration file that enters the Docker build context, eliminating the credential from build cache layers while preserving the ability to resolve release tags against the main branch.

Patch summary

In .github/workflows/docker.yml, the actions/checkout@v4 step is updated with persist-credentials: false to prevent the job token from being stored in .git/config, and fetch-depth: 0 to retrieve full history during the authenticated checkout. The subsequent git fetch origin main command in the branch-check step is removed since the origin/main reference is already available locally from the full-depth checkout.

CVSS rationale

AV:L: Exploitation requires local access to the CI/CD build environment or exported build cache; it is not remotely exploitable over a network. AC:L: Once build cache access is obtained, extracting the token from .git/config is straightforward. AT:N: No active user interaction is required. PR:H: The attacker needs high privileges (access to the CI/CD build infrastructure or build cache storage). UI:N: No user interaction is needed. VC/VI/VA:N: The CI/CD system itself is not directly compromised. SC:H: The exposed token grants full read access to the private repository, a separate system. SI:N: The job token typically has read-only scope by default; no evidence of write access. SA:N: No availability impact on the secondary system.

Weakness rationale
  • CWE-522 The GitHub Actions job token (a bearer credential) was persisted in .git/config and included in the Docker build context and cache layers without being removed, constituting insufficient protection of credentials in a build artifact.
  • CWE-200 The credential was exposed in Docker build cache layers accessible to anyone with access to the build environment, constituting information disclosure of a sensitive authentication token.
Attack pattern rationale
  • CAPEC-1 The closest plausible CAPEC is Session Hijacking: the exposed job token functions as a bearer credential that, if obtained from the build cache, allows an attacker to impersonate the CI/CD system's authenticated session and access the private repository. This is an imperfect mapping because the token is a CI/CD job token rather than a user session cookie, but the underlying mechanism (stealing a valid credential to assume an authenticated identity) aligns with the CAPEC-1 pattern. No CAPEC specifically covers credential leakage in build artifacts, so CAPEC-1 is the best available match.
Assumptions to verify
  • The GitHub Actions job token has read-only scope to the repository by default; if write scope is configured, SI could be elevated to Low.
  • The build cache is accessible to a limited set of actors (CI/CD operators or platform administrators); broader exposure (e.g., public Docker registry with cached layers) would change AV and PR ratings.
  • CAPEC-1 (Session Hijacking) is the closest available mapping; no CAPEC specifically addresses credential leakage in CI/CD build artifacts.
  • The affected version range is inferred from the explicit context (affected_version: 0, fixed_version: 4.1.0); the exact first vulnerable release is not confirmed by the patch alone.
  • The vulnerability is specific to the GitHub Actions Docker build workflow and does not affect the flowintel application runtime.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 7 11 medium 5
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-03 12:17 UTC Fix developed Corrective change authored (6b26b3f8cf9fab13244d1496443f822114ce4d05): fix: [security] Keep checkout credentials out of Docker build https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05.patch
  2. 2026-11-20 09:02 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133630",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:16:56.116043Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:17:10.037Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            ".github/workflows/docker.yml"
          ],
          "product": "flowintel",
          "programFiles": [
            ".github/workflows/docker.yml"
          ],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "The flowintel Docker build workflow used the GitHub Actions checkout action with default credential persistence, causing the job authentication token to be written into the repository\u0027s Git configuration file. This file became part of the Docker build context and could be retained in exported build cache layers, even though Git metadata was stripped from the final runtime image.\n\nAn attacker with access to the CI/CD build cache or build environment could extract the persisted token and use it to gain unauthorized access to the private source repository.\n\nImpact:\n- Exposure of a valid authentication token to the private repository\n- Potential unauthorized read access to source code and repository metadata\n\nAffected versions: flowintel \u003c 4.1.0"
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-1",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-1 Session Hijacking"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "LOCAL",
            "baseScore": 5.9,
            "baseSeverity": "MEDIUM",
            "privilegesRequired": "HIGH",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "HIGH",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:L/AC:L/AT:N/PR:H/UI:N/VC:N/VI:N/VA:N/SC:H/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "NONE",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "no"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T09:03:23Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-522",
              "description": "CWE-522 Insufficiently Protected Credentials",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-200",
              "description": "CWE-200 Exposure of Sensitive Information to an Unauthorized Actor",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T09:09:40.890Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The checkout step is configured to not persist credentials into the workspace, and the full branch history is fetched during the authenticated checkout phase. This ensures the job token is never written to the Git configuration file that enters the Docker build context, eliminating the credential from build cache layers while preserving the ability to resolve release tags against the main branch."
        }
      ],
      "title": "flowintel CI/CD Workflow Exposes Checkout Credentials in Docker Build Cache",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "The GitHub Actions job token has read-only scope to the repository by default; if write scope is configured, SI could be elevated to Low.",
                  "The build cache is accessible to a limited set of actors (CI/CD operators or platform administrators); broader exposure (e.g., public Docker registry with cached layers) would change AV and PR ratings.",
                  "CAPEC-1 (Session Hijacking) is the closest available mapping; no CAPEC specifically addresses credential leakage in CI/CD build artifacts.",
                  "The affected version range is inferred from the explicit context (affected_version: 0, fixed_version: 4.1.0); the exact first vulnerable release is not confirmed by the patch alone.",
                  "The vulnerability is specific to the GitHub Actions Docker build workflow and does not affect the flowintel application runtime."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-1",
                    "rationale": "The closest plausible CAPEC is Session Hijacking: the exposed job token functions as a bearer credential that, if obtained from the build cache, allows an attacker to impersonate the CI/CD system\u0027s authenticated session and access the private repository. This is an imperfect mapping because the token is a CI/CD job token rather than a user session cookie, but the underlying mechanism (stealing a valid credential to assume an authenticated identity) aligns with the CAPEC-1 pattern. No CAPEC specifically covers credential leakage in build artifacts, so CAPEC-1 is the best available match."
                  }
                ],
                "commit": "6b26b3f8cf9fab13244d1496443f822114ce4d05",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:L: Exploitation requires local access to the CI/CD build environment or exported build cache; it is not remotely exploitable over a network. AC:L: Once build cache access is obtained, extracting the token from .git/config is straightforward. AT:N: No active user interaction is required. PR:H: The attacker needs high privileges (access to the CI/CD build infrastructure or build cache storage). UI:N: No user interaction is needed. VC/VI/VA:N: The CI/CD system itself is not directly compromised. SC:H: The exposed token grants full read access to the private repository, a separate system. SI:N: The job token typically has read-only scope by default; no evidence of write access. SA:N: No availability impact on the secondary system.",
                "fixSummary": "The checkout step is configured to not persist credentials into the workspace, and the full branch history is fetched during the authenticated checkout phase. This ensures the job token is never written to the Git configuration file that enters the Docker build context, eliminating the credential from build cache layers while preserving the ability to resolve release tags against the main branch.",
                "generatedAt": "2026-11-20T09:03:23.235755Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 5,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 7
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "6cdbee24e7fdf8bec92a1ccaa19b9f8c7fbf5229080dfb4029456b9fe84cf493",
                "patchSummary": "In .github/workflows/docker.yml, the actions/checkout@v4 step is updated with persist-credentials: false to prevent the job token from being stored in .git/config, and fetch-depth: 0 to retrieve full history during the authenticated checkout. The subsequent git fetch origin main command in the branch-check step is removed since the origin/main reference is already available locally from the full-depth checkout.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "6b26b3f8cf9fab13244d1496443f822114ce4d05",
                    "date": "Tue, 3 Nov 2026 13:17:37 +0100",
                    "patchSha256": "6cdbee24e7fdf8bec92a1ccaa19b9f8c7fbf5229080dfb4029456b9fe84cf493",
                    "retrievedAt": "2026-11-20T09:02:44.223116Z",
                    "source": "https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05.patch",
                    "subject": "fix: [security] Keep checkout credentials out of Docker build"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "no"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T09:03:23Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Keep checkout credentials out of Docker build",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-522",
                    "rationale": "The GitHub Actions job token (a bearer credential) was persisted in .git/config and included in the Docker build context and cache layers without being removed, constituting insufficient protection of credentials in a build artifact."
                  },
                  {
                    "cweId": "CWE-200",
                    "rationale": "The credential was exposed in Docker build cache layers accessible to anyone with access to the build environment, constituting information disclosure of a sensitive authentication token."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (6b26b3f8cf9fab13244d1496443f822114ce4d05): fix: [security] Keep checkout credentials out of Docker build",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05.patch"
                    ],
                    "timestamp": "2026-11-03T12:17:37Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/6b26b3f8cf9fab13244d1496443f822114ce4d05.patch"
                    ],
                    "timestamp": "2026-11-20T09:02:44.223116Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20377"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133630",
    "datePublished": "2026-11-20T09:09:40.890Z",
    "dateReserved": "2026-11-20T09:09:38.574Z",
    "dateUpdated": "2026-11-20T18:17:10.037Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133627 (GCVE-0-2026-133627)

Vulnerability from cvelistv5 – Published: 2026-11-20 08:52 – Updated: 2026-11-20 18:16
VLAI
Title
flowintel Case History Log Injection via Line Separator Characters in User-Authored Fields
Summary
flowintel contains a log injection vulnerability in its case history (audit log) subsystem. The case history is stored as a line-delimited file where each line represents a single event attributed to a specific user. User-controlled fields, including user names, event messages, and case titles used during case merging, were written to the history file without sanitizing embedded line separator characters (such as carriage return, line feed, vertical tab, form feed, and Unicode line/paragraph separators). An authenticated user with the ability to create or modify case history entries could embed line separators in these fields. Upon writing, the injected separators would split a single intended entry into multiple lines, each of which the history format interprets as a distinct event. This allows an attacker to forge additional history entries attributed to arbitrary users, corrupting the audit trail and undermining the integrity of case records. The vulnerability affects the case history append and merge operations. No authentication bypass or privilege escalation is required beyond normal case-management access. Affected versions: flowintel prior to 4.1.0.
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 08:42 UTC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:16 UTC
CWE
  • CWE-117 - Improper Output Neutralization for Logs
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 08:42
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1.patch eab799b16980…
Confidence
medium
Commit Subject Patch SHA-256
842a948507b8 fix: [security] Keep case history entries on a single line eab799b16980…
Fix summary

The fix introduces a sanitization helper that replaces all splitlines()-recognized line separator characters (CR, LF, VT, FF, file/group/record separators, NEL, Unicode line and paragraph separators) with a single space before any user-authored content is written to the case history file. The helper is applied to the author name, the event message, and the case title during merge operations, ensuring each history entry remains on exactly one line and cannot be split into forged entries.

Patch summary

In app/case/common_core.py, a compiled regex _HISTORY_LINE_BREAKS matching all splitlines() separators is added, along with a helper function _history_one_line() that substitutes matches with a space. In save_history(), the author name and message are passed through _history_one_line() before being written. In append_history_file(), each source line and the merging case title are also sanitized via _history_one_line() before being appended to the destination history file, and an explicit newline is written to terminate each merged entry.

CVSS rationale

AV:N: The application is a web-based case management tool; input is submitted over the network. AC:L: Exploitation requires only embedding a newline character in a name, message, or title field; no race conditions or special timing. AT:N: No special attack conditions are needed. PR:L: The attacker must be an authenticated user with access to create or modify case history entries. UI:N: No victim interaction is required beyond the attacker's own submission. VC:N: No confidentiality impact; the vulnerability does not expose data. VI:L: Integrity impact is limited to forging entries in a single case's history file; it does not compromise the broader system or other cases. VA:N: No availability impact. SC/SI/SA:N: The impact is confined to the application's own log file and does not cross a security scope boundary.

Weakness rationale
  • CWE-117 User-controlled data (names, messages, titles) is written to a structured log file without neutralizing line separator characters, allowing injection of additional log entries. This is the textbook definition of CWE-117.
Attack pattern rationale
  • CAPEC-66 CAPEC-66 describes the technique of injecting CRLF or line separator characters into structured output to create additional entries or split a single record into multiple. The underlying mechanism here is identical: an attacker embeds line separators in user input to split a single log entry into multiple forged entries. The context differs (log file rather than HTTP response), but no CAPEC pattern more precisely captures structured-output injection via line separators. This is the closest available match.
Assumptions to verify
  • PR:L assumes the attacker needs at least basic authenticated access to the case management interface to submit history entries; the patch does not specify whether unauthenticated access is possible.
  • VI:L assumes the integrity impact is limited to the case history file and does not extend to other system components or databases; the patch only modifies the file-writing path.
  • CAPEC-66 is the closest available pattern for structured-output injection via line separators; no CAPEC specifically covers log-file injection, so this mapping carries some uncertainty.
  • The affected version range (0 to <4.1.0) is taken from the explicit analyst context; the patch itself does not encode version boundaries.
  • The commit message references an internal finding identifier (F124 / G29) which is not publicly resolvable and is not included as a reference.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 7 11 medium 5
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-06 09:54 UTC Fix developed Corrective change authored (842a948507b87ee8c614ba2871ffecb3172d29e1): fix: [security] Keep case history entries on a single line https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1.patch
  2. 2026-11-20 08:41 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133627",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:16:19.260199Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:16:37.656Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/case/common_core.py"
          ],
          "product": "flowintel",
          "programFiles": [
            "app/case/common_core.py"
          ],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "flowintel contains a log injection vulnerability in its case history (audit log) subsystem. The case history is stored as a line-delimited file where each line represents a single event attributed to a specific user. User-controlled fields, including user names, event messages, and case titles used during case merging, were written to the history file without sanitizing embedded line separator characters (such as carriage return, line feed, vertical tab, form feed, and Unicode line/paragraph separators).\n\nAn authenticated user with the ability to create or modify case history entries could embed line separators in these fields. Upon writing, the injected separators would split a single intended entry into multiple lines, each of which the history format interprets as a distinct event. This allows an attacker to forge additional history entries attributed to arbitrary users, corrupting the audit trail and undermining the integrity of case records.\n\nThe vulnerability affects the case history append and merge operations. No authentication bypass or privilege escalation is required beyond normal case-management access.\n\nAffected versions: flowintel prior to 4.1.0."
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-66",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-66 HTTP Request/Response Splitting"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 5.3,
            "baseSeverity": "MEDIUM",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:L/VA:N/SC:N/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "LOW",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T08:42:41Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-117",
              "description": "CWE-117 Improper Output Neutralization for Logs",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T08:52:54.104Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The fix introduces a sanitization helper that replaces all splitlines()-recognized line separator characters (CR, LF, VT, FF, file/group/record separators, NEL, Unicode line and paragraph separators) with a single space before any user-authored content is written to the case history file. The helper is applied to the author name, the event message, and the case title during merge operations, ensuring each history entry remains on exactly one line and cannot be split into forged entries."
        }
      ],
      "title": "flowintel Case History Log Injection via Line Separator Characters in User-Authored Fields",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "PR:L assumes the attacker needs at least basic authenticated access to the case management interface to submit history entries; the patch does not specify whether unauthenticated access is possible.",
                  "VI:L assumes the integrity impact is limited to the case history file and does not extend to other system components or databases; the patch only modifies the file-writing path.",
                  "CAPEC-66 is the closest available pattern for structured-output injection via line separators; no CAPEC specifically covers log-file injection, so this mapping carries some uncertainty.",
                  "The affected version range (0 to \u003c4.1.0) is taken from the explicit analyst context; the patch itself does not encode version boundaries.",
                  "The commit message references an internal finding identifier (F124 / G29) which is not publicly resolvable and is not included as a reference."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-66",
                    "rationale": "CAPEC-66 describes the technique of injecting CRLF or line separator characters into structured output to create additional entries or split a single record into multiple. The underlying mechanism here is identical: an attacker embeds line separators in user input to split a single log entry into multiple forged entries. The context differs (log file rather than HTTP response), but no CAPEC pattern more precisely captures structured-output injection via line separators. This is the closest available match."
                  }
                ],
                "commit": "842a948507b87ee8c614ba2871ffecb3172d29e1",
                "confidence": "medium",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N: The application is a web-based case management tool; input is submitted over the network. AC:L: Exploitation requires only embedding a newline character in a name, message, or title field; no race conditions or special timing. AT:N: No special attack conditions are needed. PR:L: The attacker must be an authenticated user with access to create or modify case history entries. UI:N: No victim interaction is required beyond the attacker\u0027s own submission. VC:N: No confidentiality impact; the vulnerability does not expose data. VI:L: Integrity impact is limited to forging entries in a single case\u0027s history file; it does not compromise the broader system or other cases. VA:N: No availability impact. SC/SI/SA:N: The impact is confined to the application\u0027s own log file and does not cross a security scope boundary.",
                "fixSummary": "The fix introduces a sanitization helper that replaces all splitlines()-recognized line separator characters (CR, LF, VT, FF, file/group/record separators, NEL, Unicode line and paragraph separators) with a single space before any user-authored content is written to the case history file. The helper is applied to the author name, the event message, and the case title during merge operations, ensuring each history entry remains on exactly one line and cannot be split into forged entries.",
                "generatedAt": "2026-11-20T08:42:41.961994Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 5,
                      "confidence": "medium",
                      "model": "qwen3.8:27b",
                      "score": 7
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "eab799b16980a28458755cce56ea9897af43b9a630aec94b557deb9c8f373ab4",
                "patchSummary": "In app/case/common_core.py, a compiled regex _HISTORY_LINE_BREAKS matching all splitlines() separators is added, along with a helper function _history_one_line() that substitutes matches with a space. In save_history(), the author name and message are passed through _history_one_line() before being written. In append_history_file(), each source line and the merging case title are also sanitized via _history_one_line() before being appended to the destination history file, and an explicit newline is written to terminate each merged entry.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "842a948507b87ee8c614ba2871ffecb3172d29e1",
                    "date": "Fri, 6 Nov 2026 10:54:12 +0100",
                    "patchSha256": "eab799b16980a28458755cce56ea9897af43b9a630aec94b557deb9c8f373ab4",
                    "retrievedAt": "2026-11-20T08:41:59.334828Z",
                    "source": "https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1.patch",
                    "subject": "fix: [security] Keep case history entries on a single line"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T08:42:41Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Keep case history entries on a single line",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-117",
                    "rationale": "User-controlled data (names, messages, titles) is written to a structured log file without neutralizing line separator characters, allowing injection of additional log entries. This is the textbook definition of CWE-117."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (842a948507b87ee8c614ba2871ffecb3172d29e1): fix: [security] Keep case history entries on a single line",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1.patch"
                    ],
                    "timestamp": "2026-11-06T09:54:12Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/842a948507b87ee8c614ba2871ffecb3172d29e1.patch"
                    ],
                    "timestamp": "2026-11-20T08:41:59.334828Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20212"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133627",
    "datePublished": "2026-11-20T08:52:54.104Z",
    "dateReserved": "2026-11-20T08:52:49.070Z",
    "dateUpdated": "2026-11-20T18:16:37.656Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

CVE-2026-133626 (GCVE-0-2026-133626)

Vulnerability from cvelistv5 – Published: 2026-11-20 08:52 – Updated: 2026-11-20 18:15
VLAI
Title
FlowIntel: Resource Exhaustion via Sparse CSV/JSON Table Expansion During File-to-Note Conversion
Summary
FlowIntel contains a resource exhaustion vulnerability in its file-to-note conversion feature. When a user converts an attached CSV or JSON file into a case note, the application renders the data as a Markdown table by padding every row to the maximum column count. A crafted file with a small number of filled cells but a very large number of columns (or keys) causes the application to allocate a quadratic number of padded table cells, potentially exhausting a worker process in memory or CPU. The same issue affects nested JSON values rendered inside table cells, where pretty-printed indentation further inflates the output size. Preconditions: - An authenticated user with access to a case in FlowIntel - Ability to attach and convert a CSV or JSON file to a note Impact: - Denial of service: exhaustion of a worker process, potentially disrupting service for other users - No confidentiality or integrity impact Affected versions: FlowIntel prior to 4.1.0
SSVC
Exploitation: none Automatable: yes Technical Impact: partial
Supplier · CIRCL (v2.0.3)
Decision recorded 2026-11-20 08:43 UTC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 18:15 UTC
CWE
  • CWE-400 - Uncontrolled Resource Consumption
  • CWE-770 - Allocation of Resources Without Limits or Throttling
Assigner
CIRCL CNA under the ENISA root
CNA scorecard A 92/100 over 310 records in the last 180 days details
References
Impacted products
Vendor Product Version
flowintel flowintel Affected: 0 , < 4.1.0 (semver)
    cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*
Create a notification for this product.
GCVE extensions
bcp-05-x-01
AI-assisted vulnerability information annotation
GCVE-BCP-05-X-01
Whole record AI-generated Human-reviewed GNA-1

Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.

ai-computer-assisted:llm-generatedai-computer-assisted:classification
Model Source Identifier
qwen3.8:27b ollama qwen3.8:27b
bcp-05-x-02
Patch-to-vulnerability generation provenance
GCVE-BCP-05-X-02
Generator
patch2vuln.py on 2026-11-20 08:43
Model
qwen3.8:27b
Input
https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6.patch 1aa2d65a4e08…
Confidence
high
Commit Subject Patch SHA-256
90b822a0fca3 fix: [security] Bound CSV and JSON expansion when converting 1aa2d65a4e08…
Fix summary

The fix introduces bounds on table expansion: a maximum padded-cell count (50,000) and a maximum padding ratio (4x filled cells). When a CSV or JSON dataset would exceed these limits, the original content is preserved in a safely fenced code block instead of being rendered as a table. Additionally, nested JSON values in table cells are now serialized in compact form (no indentation), and the JSON fallback path uses the same safe fencing helper, preventing both quadratic cell expansion and indentation-based output inflation.

Patch summary

In app/utils/file_converter.py: added constants MAX_PADDED_TABLE_CELLS (50000) and MAX_PADDING_RATIO (4); added helper _table_too_sparse() to detect oversized sparse tables; added helper _fenced_content() that generates a code fence long enough to prevent fence-injection from user content; modified convert_csv_to_note() to short-circuit to fenced output when the table is too sparse; modified convert_json_to_note() to apply the same sparse check for arrays of objects; changed json.dumps(value, indent=2) to json.dumps(value) for nested values in table cells; replaced the JSON fallback json.dumps(data, indent=2) with _fenced_content(content, "json"). In doc/user-manual.md: added a note documenting the new sparse-data behavior.

CVSS rationale

AV:N - the file conversion is accessible over the network via the web application. AC:L - crafting a sparse CSV/JSON file is straightforward; no race conditions or special timing required. AT:N - no prior tampering with system data needed. PR:L - requires an authenticated user with access to a case to attach and convert a file. UI:N - the conversion is triggered by the attacker's own action (uploading and converting the file); no victim interaction needed. VC:N, VI:N - no confidentiality or integrity impact. VA:H - worker exhaustion can deny service to other users of the FlowIntel instance. SC:N, SI:N, SA:N - no secondary component impact is evidenced.

Weakness rationale
  • CWE-400 The application does not bound the expansion of sparse CSV/JSON data into padded table cells, allowing a small input to trigger quadratic memory/CPU allocation and exhaust a worker. This is a classic uncontrolled resource consumption issue.
  • CWE-770 The table rendering path allocates cells proportional to rows x max_cols without any upper bound, and JSON pretty-printing further inflates output. No limit or throttle was applied before the fix.
Attack pattern rationale
  • CAPEC-10 The attacker supplies a crafted CSV or JSON file that, when processed by the conversion feature, causes the application to allocate an excessive number of table cells, exhausting worker resources. This is a direct match to application resource exhaustion via crafted input.
  • CAPEC-124 The underlying mechanism is quadratic expansion: a file with N columns and M rows produces N*M padded cells. The attacker exploits the algorithmic complexity of the table-padding logic. This is a secondary, more specific mapping; CAPEC-10 is the primary match.
Assumptions to verify
  • PR:L assumes the file-to-note conversion feature requires an authenticated user with case access; the patch does not explicitly state authentication requirements, but FlowIntel is a case-management tool where case operations are typically authenticated.
  • VA:H assumes that exhausting a single worker has a meaningful availability impact on the service; if the deployment runs many redundant workers behind a load balancer, the practical impact could be lower.
  • CAPEC-124 is included as a secondary mapping; the primary mechanism is unbounded allocation (CAPEC-10), and the quadratic complexity is the underlying cause rather than a separate attack pattern.
  • The affected version range is taken from the explicit context (affected_version 0, fixed_version 4.1.0); the patch itself does not encode version boundaries.
  • The _fenced_content helper also mitigates a potential Markdown fence-injection issue (a crafted backtick sequence in the file content could break out of the code block), but the primary vulnerability addressed is resource exhaustion.
Model comparison

Selected qwen3.8:27b by deterministic-consensus-v1
The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required.

Model Score Agreement Confidence Assumptions
qwen3.8:27b 8 11 high 5
bcp-05-x-03
Vulnerability handling and disclosure timeline
GCVE-BCP-05-X-03
  1. 2026-11-06 09:53 UTC Fix developed Corrective change authored (90b822a0fca35d33688730772b964b478c7fcde6): fix: [security] Bound CSV and JSON expansion when converting https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6.patch
  2. 2026-11-20 08:42 UTC Fix available Patch retrieved from the supplied URL; retrieval time is used as the default availability time. https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6.patch
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133626",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T18:15:29.296559Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T18:15:48.121Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "cpes": [
            "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*"
          ],
          "modules": [
            "app/utils/file_converter.py"
          ],
          "product": "flowintel",
          "programFiles": [
            "app/utils/file_converter.py"
          ],
          "repo": "https://github.com/flowintel/flowintel",
          "vendor": "flowintel",
          "versions": [
            {
              "lessThan": "4.1.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:*",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "elhoim"
        },
        {
          "lang": "en",
          "type": "remediation developer",
          "value": "David Cruciani"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "FlowIntel contains a resource exhaustion vulnerability in its file-to-note conversion feature. When a user converts an attached CSV or JSON file into a case note, the application renders the data as a Markdown table by padding every row to the maximum column count. A crafted file with a small number of filled cells but a very large number of columns (or keys) causes the application to allocate a quadratic number of padded table cells, potentially exhausting a worker process in memory or CPU.\n\nThe same issue affects nested JSON values rendered inside table cells, where pretty-printed indentation further inflates the output size.\n\nPreconditions:\n- An authenticated user with access to a case in FlowIntel\n- Ability to attach and convert a CSV or JSON file to a note\n\nImpact:\n- Denial of service: exhaustion of a worker process, potentially disrupting service for other users\n- No confidentiality or integrity impact\n\nAffected versions: FlowIntel prior to 4.1.0"
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-10",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-10 Application Resource Exhaustion"
            }
          ]
        },
        {
          "capecId": "CAPEC-124",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-124 Resource Exhaustion via Algorithmic Complexity"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "NETWORK",
            "baseScore": 7.1,
            "baseSeverity": "HIGH",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "NONE",
            "subIntegrityImpact": "NONE",
            "userInteraction": "NONE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "HIGH",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "NONE",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        },
        {
          "format": "SSVC",
          "other": {
            "content": {
              "options": [
                {
                  "Exploitation": "none"
                },
                {
                  "Automatable": "yes"
                },
                {
                  "Technical Impact": "partial"
                }
              ],
              "role": "Supplier",
              "timestamp": "2026-11-20T08:43:07Z",
              "version": "2.0.3"
            },
            "type": "SSVC"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-400",
              "description": "CWE-400 Uncontrolled Resource Consumption",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-770",
              "description": "CWE-770 Allocation of Resources Without Limits or Throttling",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T08:52:28.073Z",
        "orgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
        "shortName": "CIRCL"
      },
      "references": [
        {
          "name": "Security patch",
          "tags": [
            "patch"
          ],
          "url": "https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6"
        }
      ],
      "solutions": [
        {
          "lang": "en",
          "value": "The fix introduces bounds on table expansion: a maximum padded-cell count (50,000) and a maximum padding ratio (4x filled cells). When a CSV or JSON dataset would exceed these limits, the original content is preserved in a safely fenced code block instead of being rendered as a table. Additionally, nested JSON values in table cells are now serialized in compact form (no indentation), and the JSON fallback path uses the same safe fencing helper, preventing both quadratic cell expansion and indentation-based output inflation."
        }
      ],
      "title": "FlowIntel: Resource Exhaustion via Sparse CSV/JSON Table Expansion During File-to-Note Conversion",
      "x_gcve": [
        {
          "extensions": {
            "bcp-05-x-01": {
              "ai_annotations": [
                {
                  "ai_level": "generated",
                  "description": "Draft vulnerability metadata was generated from a git-format patch using an Ollama-hosted language model. Human validation is required before publication.",
                  "gna_source": 1,
                  "models": [
                    {
                      "gna_source": 1,
                      "identifier": "qwen3.8:27b",
                      "name": "qwen3.8:27b",
                      "source": "ollama"
                    }
                  ],
                  "review_status": "full",
                  "scope": "record",
                  "tags": [
                    "ai-computer-assisted:llm-generated",
                    "ai-computer-assisted:classification"
                  ]
                }
              ]
            },
            "bcp-05-x-02": {
              "x_patch2vuln": {
                "assumptions": [
                  "PR:L assumes the file-to-note conversion feature requires an authenticated user with case access; the patch does not explicitly state authentication requirements, but FlowIntel is a case-management tool where case operations are typically authenticated.",
                  "VA:H assumes that exhausting a single worker has a meaningful availability impact on the service; if the deployment runs many redundant workers behind a load balancer, the practical impact could be lower.",
                  "CAPEC-124 is included as a secondary mapping; the primary mechanism is unbounded allocation (CAPEC-10), and the quadratic complexity is the underlying cause rather than a separate attack pattern.",
                  "The affected version range is taken from the explicit context (affected_version 0, fixed_version 4.1.0); the patch itself does not encode version boundaries.",
                  "The _fenced_content helper also mitigates a potential Markdown fence-injection issue (a crafted backtick sequence in the file content could break out of the code block), but the primary vulnerability addressed is resource exhaustion."
                ],
                "capecRationale": [
                  {
                    "capecId": "CAPEC-10",
                    "rationale": "The attacker supplies a crafted CSV or JSON file that, when processed by the conversion feature, causes the application to allocate an excessive number of table cells, exhausting worker resources. This is a direct match to application resource exhaustion via crafted input."
                  },
                  {
                    "capecId": "CAPEC-124",
                    "rationale": "The underlying mechanism is quadratic expansion: a file with N columns and M rows produces N*M padded cells. The attacker exploits the algorithmic complexity of the table-padding logic. This is a secondary, more specific mapping; CAPEC-10 is the primary match."
                  }
                ],
                "commit": "90b822a0fca35d33688730772b964b478c7fcde6",
                "confidence": "high",
                "credits": [
                  {
                    "lang": "en",
                    "type": "reporter",
                    "value": "elhoim"
                  },
                  {
                    "lang": "en",
                    "type": "remediation developer",
                    "value": "David Cruciani"
                  }
                ],
                "cvssRationale": "AV:N - the file conversion is accessible over the network via the web application. AC:L - crafting a sparse CSV/JSON file is straightforward; no race conditions or special timing required. AT:N - no prior tampering with system data needed. PR:L - requires an authenticated user with access to a case to attach and convert a file. UI:N - the conversion is triggered by the attacker\u0027s own action (uploading and converting the file); no victim interaction needed. VC:N, VI:N - no confidentiality or integrity impact. VA:H - worker exhaustion can deny service to other users of the FlowIntel instance. SC:N, SI:N, SA:N - no secondary component impact is evidenced.",
                "fixSummary": "The fix introduces bounds on table expansion: a maximum padded-cell count (50,000) and a maximum padding ratio (4x filled cells). When a CSV or JSON dataset would exceed these limits, the original content is preserved in a safely fenced code block instead of being rendered as a table. Additionally, nested JSON values in table cells are now serialized in compact form (no indentation), and the JSON fallback path uses the same safe fencing helper, preventing both quadratic cell expansion and indentation-based output inflation.",
                "generatedAt": "2026-11-20T08:43:07.422130Z",
                "generator": "patch2vuln.py",
                "model": "qwen3.8:27b",
                "modelComparison": {
                  "potentialErrors": [],
                  "rankings": [
                    {
                      "agreementScore": 11,
                      "assumptionCount": 5,
                      "confidence": "high",
                      "model": "qwen3.8:27b",
                      "score": 8
                    }
                  ],
                  "selectedModel": "qwen3.8:27b",
                  "selectionMethod": "deterministic-consensus-v1",
                  "selectionNotice": "The selected result is closest to model consensus; this heuristic does not establish factual correctness and human review remains required."
                },
                "patchSha256": "1aa2d65a4e08d2c447aea812aa20e4190cf65a5be1e6a751afaae121545e51f4",
                "patchSummary": "In app/utils/file_converter.py: added constants MAX_PADDED_TABLE_CELLS (50000) and MAX_PADDING_RATIO (4); added helper _table_too_sparse() to detect oversized sparse tables; added helper _fenced_content() that generates a code fence long enough to prevent fence-injection from user content; modified convert_csv_to_note() to short-circuit to fenced output when the table is too sparse; modified convert_json_to_note() to apply the same sparse check for arrays of objects; changed json.dumps(value, indent=2) to json.dumps(value) for nested values in table cells; replaced the JSON fallback json.dumps(data, indent=2) with _fenced_content(content, \"json\"). In doc/user-manual.md: added a note documenting the new sparse-data behavior.",
                "patchTruncated": false,
                "patches": [
                  {
                    "commit": "90b822a0fca35d33688730772b964b478c7fcde6",
                    "date": "Fri, 6 Nov 2026 10:53:20 +0100",
                    "patchSha256": "1aa2d65a4e08d2c447aea812aa20e4190cf65a5be1e6a751afaae121545e51f4",
                    "retrievedAt": "2026-11-20T08:42:42.445428Z",
                    "source": "https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6.patch",
                    "sourceUrl": "https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6.patch",
                    "subject": "fix: [security] Bound CSV and JSON expansion when converting"
                  }
                ],
                "source": "https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6.patch",
                "ssvc": {
                  "options": [
                    {
                      "Exploitation": "none"
                    },
                    {
                      "Automatable": "yes"
                    },
                    {
                      "Technical Impact": "partial"
                    }
                  ],
                  "role": "Supplier",
                  "timestamp": "2026-11-20T08:43:07Z",
                  "version": "2.0.3"
                },
                "subject": "fix: [security] Bound CSV and JSON expansion when converting",
                "weaknessRationale": [
                  {
                    "cweId": "CWE-400",
                    "rationale": "The application does not bound the expansion of sparse CSV/JSON data into padded table cells, allowing a small input to trigger quadratic memory/CPU allocation and exhaust a worker. This is a classic uncontrolled resource consumption issue."
                  },
                  {
                    "cweId": "CWE-770",
                    "rationale": "The table rendering path allocates cells proportional to rows x max_cols without any upper bound, and JSON pretty-printing further inflates output. No limit or throttle was applied before the fix."
                  }
                ]
              }
            },
            "bcp-05-x-03": {
              "x_timeline": {
                "events": [
                  {
                    "description": "Corrective change authored (90b822a0fca35d33688730772b964b478c7fcde6): fix: [security] Bound CSV and JSON expansion when converting",
                    "id": "evt-fix-developed-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6.patch"
                    ],
                    "timestamp": "2026-11-06T09:53:20Z",
                    "type": "fix-developed"
                  },
                  {
                    "description": "Patch retrieved from the supplied URL; retrieval time is used as the default availability time.",
                    "id": "evt-fix-available-1",
                    "references": [
                      "https://github.com/flowintel/flowintel/commit/90b822a0fca35d33688730772b964b478c7fcde6.patch"
                    ],
                    "timestamp": "2026-11-20T08:42:42.445428Z",
                    "type": "fix-available"
                  }
                ]
              }
            }
          },
          "recordType": "advisory",
          "vulnId": "GCVE-1-2026-20366"
        }
      ]
    }
  },
  "cveMetadata": {
    "assignerOrgId": "5a6e4751-2f3f-4070-9419-94fb35b644e8",
    "assignerShortName": "CIRCL",
    "cveId": "CVE-2026-133626",
    "datePublished": "2026-11-20T08:52:28.073Z",
    "dateReserved": "2026-11-20T08:52:20.420Z",
    "dateUpdated": "2026-11-20T18:15:48.121Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}
displaying 1 - 10 assigned vulnerabilities in total 315