<?xml version='1.0' encoding='UTF-8'?>
<?xml-stylesheet href="/static/style.xsl" type="text/xsl"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="en">
  <id>https://vulnerability.circl.lu/rss/recent/osv_haskell/10</id>
  <title>Most recent entries from osv_haskell</title>
  <updated>2026-11-22T14:03:46.368482+00:00</updated>
  <author>
    <name>Vulnerability-Lookup</name>
    <email>info@circl.lu</email>
  </author>
  <link href="https://vulnerability.circl.lu" rel="alternate"/>
  <generator uri="https://lkiesow.github.io/python-feedgen" version="1.0.0">python-feedgen</generator>
  <subtitle>Contains only the most 10 recent entries.</subtitle>
  <entry>
    <id>https://vulnerability.circl.lu/vuln/hsec-2026-0016</id>
    <title>HSEC-2026-0016 — Improper IP address validation in validateCertificateName</title>
    <updated>2026-11-17T09:58:05+00:00</updated>
    <content type="xhtml">
      <div xmlns="http://www.w3.org/1999/xhtml"><p><strong>Affected:</strong> Hackage: crypton-x509-validation</p>
<p># Improper IP address validation in validateCertificateName</p>
<p>`Data.X509.Validation.getNames` extracts only `AltNameDNS` entries from the
subjectAltName extension (Data/X509/Validation.hs, `unAltName`,
~L512–515 in crypton-x509-validation 1.10.1). When a certificate carries an
`iPAddress` SAN but no `dNSName` SAN, the extracted name list is empty, so
`validateNames` (~L569–580) falls through to the commonName branch as if no
SAN extension existed, even though the extension's haddock promises that the
commonName is ignored whenever subjectAltName is present. Consequently, when
a client connects to an IP literal:</p>
<p>* a certificate with `CN = &lt;target IP&gt;` but `iPAddress` SAN pointing to a
  different host is **accepted**;
* a certificate whose only valid identity is a matching `iPAddress` SAN is
  **rejected**.</p>
<p>Exploitation requires a certificate from a CA the client trusts that has a
commonName different from its iPAddress SAN (public CAs are barred from this
by the CA/Browser Forum Baseline Requirements, but private or enterprise
CAs can issue such certificates). A successful impersonation exposes and can
modify the protected connection.</p>
<p>The draft fix makes `getNames` distinguish "no SAN extension" from "SAN
extension present with no DNS names" and applies SAN semantics (including
the existing iPAddress comparison) whenever the extension is present,
restoring the documented SAN-supersedes-CN behavior in line with RFC 9525.
No fixed release is published at the time of writing…</p></div>
    </content>
    <link href="https://vulnerability.circl.lu/vuln/hsec-2026-0016"/>
    <published>2026-11-17T09:58:05+00:00</published>
  </entry>
  <entry>
    <id>https://vulnerability.circl.lu/vuln/hsec-2026-0017</id>
    <title>HSEC-2026-0017 — Out-of-bounds writes in BMP RLE and TIFF strip decoding</title>
    <updated>2026-11-17T09:55:51+00:00</updated>
    <content type="xhtml">
      <div xmlns="http://www.w3.org/1999/xhtml"><p><strong>Affected:</strong> Hackage: JuicyPixels</p>
<p># Out-of-bounds writes in BMP RLE and TIFF strip decoding</p>
<p>JuicyPixels contains independent out-of-bounds writes reachable from ordinary
image decoding (`Codec.Picture.decodeImage` and the format-specific decoders):</p>
<p>1. **BMP `BI_RLE8`/`BI_RLE4`** (`Codec.Picture.Bitmap.decodeImageY8RLE`): the
   pixel buffer size is computed with 32-bit `Int32` arithmetic and wraps to a
   zero-length allocation for large headers; end-of-line and vertical-delta
   control codes move the destination cursor without a bound, and the
   horizontal delta jump is unclamped. Every run writes attacker-chosen bytes
   through `M.unsafeWrite` at the cursor position — negative offsets before
   the raster, arbitrary offsets against a wrapped allocation, and up to 255
   bytes past it.
2. **TIFF uncompressed/PackBits contiguous strips**
   (`Codec.Picture.Tiff.gatherStrips`, `copyByteString`, `unpackPackBit`):
   strip copies have no destination bound and the per-strip size computation
   wraps negative for strips starting past the raster, so excess strips copy
   attacker-chosen bytes past the end of the allocated raster.</p>
<p>Impact: heap memory adjacent to the decoded image is corrupted with
attacker-selected values and offsets through maliciously crafted BMP or TIFF
files. Beyond corruption, exploitation primitives were demonstrated
end-to-end on GHC 9.6 / linux-amd64: a crafted BMP deterministically crashed
the process via the wrapped allocation; against a warp-based service a single
uploaded BMP over…</p></div>
    </content>
    <link href="https://vulnerability.circl.lu/vuln/hsec-2026-0017"/>
    <published>2026-11-17T09:55:51+00:00</published>
  </entry>
  <entry>
    <id>https://vulnerability.circl.lu/vuln/hsec-2026-0015</id>
    <title>HSEC-2026-0015 — Missing validation of CONNECT requests in http2</title>
    <updated>2026-11-16T02:22:07+00:00</updated>
    <content type="xhtml">
      <div xmlns="http://www.w3.org/1999/xhtml"><p><strong>Affected:</strong> Hackage: http2</p>
<p># Missing validation of CONNECT requests in http2</p>
<p>For CONNECT requests, `checkRequestHeader` (`Network/HTTP2/H2/HPACK.hs`) in http2 up to 5.4.13 returned early, only requiring the absence of `:scheme` and `:path`. Every other check was skipped:</p>
<p>* the requirement that `:authority` is present (RFC 9113, Section 8.5);
* the check that `:authority` and `Host` agree, which applies to every other request (RFC 9113, Section 8.3.1: a server SHOULD treat a mismatch as malformed);
* the common checks for `:status`, `Connection` and `TE`.</p>
<p>A CONNECT request carrying `:authority: a.example` and `Host: b.example` was therefore accepted. Consumers that read `Host` see `b.example` where intermediaries that read `:authority` see `a.example`. warp, for example, sets `requestHeaderHost` to `Host` when present and falls back to `:authority`, so an application behind an intermediary that routes or authorizes on `:authority` handles a request other than the one the intermediary validated — a request can be smuggled past the front-end's access control, validation, logging or WAF rules.</p>
<p>Exploitation requires a front-end that makes decisions based on `:authority` while the application behind it bases routing or authorization on `Host`; a directly exposed server sees only one authority.</p>
<p>A CONNECT request with neither `:authority` nor `:path` also passed validation and reached warp 3.4.21, where `fromJust (mPath &lt;|&gt; mAuth)` (`Network/Wai/Handler/Warp/HTTP2/Request.hs`) threw. The exception was ra…</p></div>
    </content>
    <link href="https://vulnerability.circl.lu/vuln/hsec-2026-0015"/>
    <published>2026-11-16T02:22:07+00:00</published>
  </entry>
  <entry>
    <id>https://vulnerability.circl.lu/vuln/hsec-2026-0014</id>
    <title>HSEC-2026-0014 — Arbitrary file write via path traversal in unpackToDir</title>
    <updated>2026-11-13T11:01:54+00:00</updated>
    <content type="xhtml">
      <div xmlns="http://www.w3.org/1999/xhtml"><p><strong>Affected:</strong> Hackage: libarchive</p>
<p># Arbitrary file write via path traversal in unpackToDir</p>
<p>`Codec.Archive.unpackToDir` extracts a tar (or other archive supported by the
libarchive C library) from a `ByteString` into a directory. `unpackToDirLazy`
is the streaming variant; both funnel through the same internal
`unpackEntriesFp` routine.</p>
<p>`unpackEntriesFp` joined each entry's name onto the destination directory with
`System.FilePath.(&lt;/&gt;)` without validating the name. Entry names starting with
`..` sequences therefore escape the destination: `fp &lt;/&gt; "../../../../etc/x"`
resolves outside the extraction root. The joined path is written back into the
archive entry (`archiveEntrySetPathname`) and handed to libarchive's
`archive_read_extract` with no `ARCHIVE_EXTRACT_SECURE_*` flags, so neither the
Haskell layer nor the C library rejects the traversal. The `..` check is not
limited to tar archives: any format whose entry names libarchive reports
verbatim is affected. Extraction result codes were also discarded, so write
failures could pass unnoticed.</p>
<p>An attacker who can supply an archive to a service that extracts it with these
APIs gains an arbitrary file write with the privileges of the extracting
process. A demonstrated escalation (verified in a container): an archive with
member `../../../../etc/profile.d/zz-poc-zipslip.sh` plants a shell script
sourced by every subsequent login shell, converting the file write into code
execution when any user logs in.</p>
<p>The reported traversal is fixed in libarchive-3.0.6.0,…</p></div>
    </content>
    <link href="https://vulnerability.circl.lu/vuln/hsec-2026-0014"/>
    <published>2026-11-13T11:01:54+00:00</published>
  </entry>
  <entry>
    <id>https://vulnerability.circl.lu/vuln/hsec-2026-0013</id>
    <title>HSEC-2026-0013 — HTTP Request Smuggling in warp</title>
    <updated>2026-11-04T12:43:59+00:00</updated>
    <content type="xhtml">
      <div xmlns="http://www.w3.org/1999/xhtml"><p><strong>Affected:</strong> Hackage: warp</p>
<p># HTTP Request Smuggling in warp</p>
<p>warp's HTTP/1.x request parser accepted request framing that RFC 9112 requires recipients to refuse, so a request could frame one way for a front-end proxy reading it and another way for warp behind it. Whoever reads such a request the other way sees one request where the other sees two, which smuggles a request past the proxy: past access control, request validation, logging, or WAF rules. warp's HTTP/2 support delegates parsing to the `http2` package and is not affected.</p>
<p>Six defects were fixed in warp 3.4.17:</p>
<p>* `Transfer-Encoding` and `Content-Length` field lines were collapsed to a single value, the last line winning. A second `Content-Length` could hide the first, and a second `Transfer-Encoding` could hide a `chunked` coding.
* Sending `Content-Length` together with `Transfer-Encoding`, or a `Transfer-Encoding` whose final coding is not `chunked`, was not refused; warp resolved the ambiguity by reading whichever framing it recognised.
* `Content-Length` values were parsed by reading the leading digits: a non-numeric value, or one too large for an `Int` which then wrapped, became length 0 — so the request body was read as the next request on the connection.
* Header field names were not validated: an obs-fold continuation line, a name containing whitespace or control characters, or an empty name became a header warp did not act on but a repairing intermediary could, including a hidden `Content-Length`.
* A bare CR or a NUL inside a hea…</p></div>
    </content>
    <link href="https://vulnerability.circl.lu/vuln/hsec-2026-0013"/>
    <published>2026-11-04T12:43:59+00:00</published>
  </entry>
  <entry>
    <id>https://vulnerability.circl.lu/vuln/hsec-2026-0012</id>
    <title>HSEC-2026-0012 — Out-of-bounds read in Eq and Ord on numbers with extreme exponents</title>
    <updated>2026-11-03T22:32:58+00:00</updated>
    <content type="xhtml">
      <div xmlns="http://www.w3.org/1999/xhtml"><p><strong>Affected:</strong> Hackage: scientific</p>
<p># Out-of-bounds read in Eq and Ord on numbers with extreme exponents</p>
<p>`scientific-0.3.9.0` reimplemented `Eq` and `Ord` without normalising their
arguments. The new implementation short-circuits on
`integerLog10' c + e` computed in `Int`, which overflows when the base-10
exponent is close to `maxBound :: Int`. The overflowed comparison then falls
through to `magnitude (e1 - e2)`, where the exponent difference overflows to a
negative `Int` as well. `magnitude` indexed its cached power-of-ten array
without a bounds check, so the negative index performed an out-of-bounds read
and the process died with `SIGSEGV`.</p>
<p>With `scientific-0.3.9.0` (GHC 9.10.3, x86_64-linux), as reported in
[basvandijk/scientific#120](https://github.com/basvandijk/scientific/issues/120):</p>
<p>```haskell
import Data.Scientific</p>
<p>main :: IO ()
main = print (scientific 10 maxBound == scientific 1 minBound)
```</p>
<p>```
$ ./repro; echo "exit=$?"
exit=139
```</p>
<p>The issue was discovered when Stackage nightly test suites
(`genvalidity-hspec-aeson`, `genvalidity-sydtest-aeson`) started segfaulting
after upgrading to `scientific-0.3.9.0`: their property tests round-trip
`aeson` `Value`s, which produces `Scientific`s with extreme exponents.</p>
<p>Versions before `0.3.9.0` did not crash, but `Eq`/`Ord` returned invalid
results there instead (the normalisation-based implementation overflows the
exponent, reporting `scientific 10 maxBound == scientific 1 minBound` as
`True`). The same unchecked read was also reachable through `(+)`…</p></div>
    </content>
    <link href="https://vulnerability.circl.lu/vuln/hsec-2026-0012"/>
    <published>2026-11-03T22:32:58+00:00</published>
  </entry>
  <entry>
    <id>https://vulnerability.circl.lu/vuln/hsec-2026-0011</id>
    <title>HSEC-2026-0011 — Superlinear encoding of numbers with large coefficients</title>
    <updated>2026-11-03T22:32:58+00:00</updated>
    <content type="xhtml">
      <div xmlns="http://www.w3.org/1999/xhtml"><p><strong>Affected:</strong> Hackage: scientific</p>
<p># Superlinear encoding of numbers with large coefficients</p>
<p>Encoding (pretty-printing) a `Scientific` number whose coefficient has many
digits took time quadratic in the number of digits: the decimal digits were
extracted by repeatedly dividing the whole coefficient by 10, once per digit.
The `Show` instance, `Data.Text.Lazy.Builder.Scientific` and
`Data.ByteString.Builder.Scientific` — which `aeson` uses to encode JSON
numbers — all share this code path. A number with one million digits took
about a minute to show, and about a tenth of a second after the fix.</p>
<p>This is a denial-of-service risk for applications that decode numbers from an
untrusted source and encode them again, e.g. JSON proxies, loggers or API
gateways built on `aeson`.</p>
<p>## Resolution</p>
<p>`scientific-0.3.9.0` reimplemented `normalize` and `toDecimalDigits` (the
digit-extraction primitive) to be practically linear in the coefficient size.</p>
<p>Note that `scientific-0.3.9.0` introduced a regression of its own; see
HSEC-2026-0012. Upgrade
directly to `scientific-0.3.10.0` or later, which fixes both issues.</p>
<p>## Credits</p>
<p>Reported by Andrzej Rybczak (@arybczak); fixed by Oleg Grenrus (@phadej).</p></div>
    </content>
    <link href="https://vulnerability.circl.lu/vuln/hsec-2026-0011"/>
    <published>2026-11-03T22:32:58+00:00</published>
  </entry>
  <entry>
    <id>https://vulnerability.circl.lu/vuln/hsec-2026-0010</id>
    <title>HSEC-2026-0010 — Improper parsing of fractional NumericDate from JWT exp claims</title>
    <updated>2026-09-29T14:22:37+00:00</updated>
    <content type="xhtml">
      <div xmlns="http://www.w3.org/1999/xhtml"><p><strong>Affected:</strong> Hackage: jwt</p>
<p># Improper parsing of fractional `NumericDate` from JWT `exp` claims</p>
<p>Since its [initial commit][], package jwt uses [`Data.Scientific.coefficient`][scientific] to extract the fractional mantissa in a way that improperly *discards the exponent*. This can lead to incorrect parsing of the `exp` claim in JWT payload body.</p>
<p>PoC:</p>
<p>```haskell
cabal repl --build-depends jwt==0.11.0</p>
<p>import Web.JWT
:set -XOverloadedStrings</p>
<p>let signer = VerifyHMACSecret "dummy-signing-symmetric-hmac-key.not-so-secret"
let exploit = "eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJzdWIiOiIxMjM0NTY3ODkwIiwibmFtZSI6IkhvbmVzdCBKb2UiLCJhZG1pbiI6dHJ1ZSwiaWF0IjoxNzkwMTY2Nzg2LCJleHAiOjE3OTAxNzAzODYuMDAwMX0.xQyl3rBYtVTxPL_VXSawSSmZkuF4MzOkzwdQGJFjE2o"
-- {
--   "sub": "1234567890",
--   "name": "Honest Joe",
--   "admin": true,
--   "iat": 1790166786,
--   "exp": 1790170386.0001
-- }
λλ ➔ Just whoops = Web.JWT.decodeAndVerifySignature signer exploit
λ
λλ ➔ whoops
Verified (JOSEHeader {typ = Just "JWT", cty = Nothing, alg = Just HS256, kid = Nothing}) (JWTClaimsSet {iss = Nothing, sub = Just 1234567890, aud = Nothing, exp = Just (NumericDate 17901703860001), nbf = Nothing, iat = Just (NumericDate 1790166786), jti = Nothing, unregisteredClaims = ClaimsMap {unClaimsMap = fromList [("admin",Bool True),("name",String "Honest Joe")]}}) (Signature "xQyl3rBYtVTxPL_VXSawSSmZkuF4MzOkzwdQGJFjE2o")
λ
λλ ➔ fmap secondsSinceEpoch . Web.JWT.iat . claims $ whoops
Just 1790166786s
λ
λλ ➔ fmap secondsSinceEpoch . Web.JWT.exp . claims…</p></div>
    </content>
    <link href="https://vulnerability.circl.lu/vuln/hsec-2026-0010"/>
    <published>2026-09-29T14:22:37+00:00</published>
  </entry>
  <entry>
    <id>https://vulnerability.circl.lu/vuln/hsec-2026-0008</id>
    <title>HSEC-2026-0008 — crypton-x509-validation and crypton-x509 do not enforce X.509 Name Constraints</title>
    <updated>2026-06-03T13:30:48+00:00</updated>
    <content type="xhtml">
      <div xmlns="http://www.w3.org/1999/xhtml"><p><strong>Affected:</strong> Hackage: crypton-x509-validation, Hackage: crypton-x509, Hackage: x509-validation, Hackage: x509</p>
<p># crypton-x509-validation and crypton-x509 do not enforce X.509 Name Constraints</p>
<p>The `crypton-x509-validation` and `crypton-x509` libraries did not
enforce the X.509 Name Constraints extension during certificate
validation. The Name Constraints extension is a critical X.509
extension that restricts the namespace (permitted and excluded
subtrees) for which a CA is authorized to issue certificates.</p>
<p>Without this enforcement, a TLS client would accept certificates with
Subject Alternative Names (SANs) that fall outside the issuing CA's
permitted subtrees. An attacker with access to a name-constrained
sub-CA's private key could therefore issue certificates for domains
outside the sub-CA's intended scope, enabling impersonation of
arbitrary domains and man-in-the-middle attacks on TLS connections.</p>
<p>The older `x509` and `x509-validation` packages are also affected but
are no longer maintained and have no fix available.</p>
<p>This issue was fixed in `crypton-x509-validation-1.9.1` and
`crypton-x509-1.9.1`.</p></div>
    </content>
    <link href="https://vulnerability.circl.lu/vuln/hsec-2026-0008"/>
    <published>2026-06-03T13:30:48+00:00</published>
  </entry>
  <entry>
    <id>https://vulnerability.circl.lu/vuln/hsec-2026-0007</id>
    <title>HSEC-2026-0007 — Denial of Service and Memory Exhaustion in aeson</title>
    <updated>2026-05-22T07:02:58+00:00</updated>
    <content type="xhtml">
      <div xmlns="http://www.w3.org/1999/xhtml"><p><strong>Affected:</strong> Hackage: aeson, Hackage: text-iso8601</p>
<p># Denial of Service and Memory Exhaustion in aeson</p>
<p>A Denial of Service (DoS) and memory exhaustion vulnerability was identified in the `aeson` package. The vulnerability allows an attacker to exhaust server memory and crash the host process by supplying maliciously crafted JSON payloads.</p>
<p>The vulnerability exists in `aeson`'s `withBoundedScientific_` function (located in `src/Data/Aeson/Types/FromJSON.hs`). The exponent bounds check only rejects large positive exponents (`exp10 &gt; 1024`) but fails to reject arbitrarily large negative exponents.</p>
<p>When an attacker sends a JSON number with a massive negative exponent (e.g., `1e-999999999`), the value bypasses the check and flows into `realToFrac`, which computes `fromRational . toRational`. For such a large negative exponent, `toRational` produces a GMP Integer with approximately 1 billion decimal digits, causing immediate and severe memory exhaustion.</p>
<p>Affected `FromJSON` instances:</p>
<p>* `Fixed a` (including `Centi`, `Pico`, `Nano`, etc.)
* `NominalDiffTime`
* `DiffTime`</p>
<p>## Resolution</p>
<p>The issue was resolved by introducing proper bounds checks:</p>
<p>- `aeson` now applies an absolute bounds check to both positive and negative exponents (`abs exp10 &gt; 1024`).</p>
<p>The fix first shipped in `aeson-2.3.0.0`, and have been backported to the previous release series as `aeson-2.2.5.1`.</p>
<p>Users are strongly advised to update to the patched versions:</p>
<p>* `aeson-2.2.5.1` or later</p>
<p>## Acknowledgements</p>
<p>The vulnerabilities were reported Nathan Walsh,…</p></div>
    </content>
    <link href="https://vulnerability.circl.lu/vuln/hsec-2026-0007"/>
    <published>2026-05-22T07:02:58+00:00</published>
  </entry>
</feed>
