<?xml version='1.0' encoding='UTF-8'?>
<?xml-stylesheet href="/static/style.xsl" type="text/xsl"?>
<rss xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/" version="2.0">
  <channel>
    <title>Most recent entries from osv_rustsec</title>
    <link>https://vulnerability.circl.lu</link>
    <description>Contains only the most 10 recent entries.</description>
    <docs>http://www.rssboard.org/rss-specification</docs>
    <generator>python-feedgen</generator>
    <language>en</language>
    <lastBuildDate>Sun, 22 Nov 2026 14:22:33 +0000</lastBuildDate>
    <item>
      <title>RUSTSEC-2026-0592 — Stack overflow in `Url::origin()` on nested `blob:` URLs</title>
      <link>https://vulnerability.circl.lu/vuln/rustsec-2026-0592</link>
      <description>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: url&lt;/p&gt;
&lt;p&gt;In `url` 2.5.8 and earlier, `Url::origin()` handled a `blob:` URL by parsing its path as another URL and calling itself on the result. It did so regardless of the inner URL&amp;#39;s scheme, `blob` included. Each `blob:` prefix cost one stack frame, with no limit on nesting depth. `Url::parse()` accepts a URL with 10,000 `blob:` prefixes followed by `https://example.com/` (50,020 bytes in total); calling `Url::origin()` on it overflowed the default 2 MiB stack of a spawned thread.&lt;/p&gt;
&lt;p&gt;A stack overflow aborts the whole process, and `catch_unwind()` does not catch it. One such URL crashes any program that calls `Url::origin()` on untrusted input. `url::quirks::origin()` also calls `Url::origin()`. For example, Servo&amp;#39;s `URL.origin` DOM getter calls `url::quirks::origin()`, so a script on a page can trigger the overflow.&lt;/p&gt;
&lt;p&gt;`url` 2.5.9 follows the URL Standard: `Url::origin()` takes the origin of the inner URL only when its scheme is `http`, `https`, or `file`, returns an opaque origin otherwise, and never calls itself. A `blob:` URL with another inner scheme that previously produced a tuple origin, such as `blob:ws://...` or `blob:blob:https://...`, now returns an opaque origin.&lt;/p&gt;</description>
      <content:encoded>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: url&lt;/p&gt;
&lt;p&gt;In `url` 2.5.8 and earlier, `Url::origin()` handled a `blob:` URL by parsing its path as another URL and calling itself on the result. It did so regardless of the inner URL&amp;#39;s scheme, `blob` included. Each `blob:` prefix cost one stack frame, with no limit on nesting depth. `Url::parse()` accepts a URL with 10,000 `blob:` prefixes followed by `https://example.com/` (50,020 bytes in total); calling `Url::origin()` on it overflowed the default 2 MiB stack of a spawned thread.&lt;/p&gt;
&lt;p&gt;A stack overflow aborts the whole process, and `catch_unwind()` does not catch it. One such URL crashes any program that calls `Url::origin()` on untrusted input. `url::quirks::origin()` also calls `Url::origin()`. For example, Servo&amp;#39;s `URL.origin` DOM getter calls `url::quirks::origin()`, so a script on a page can trigger the overflow.&lt;/p&gt;
&lt;p&gt;`url` 2.5.9 follows the URL Standard: `Url::origin()` takes the origin of the inner URL only when its scheme is `http`, `https`, or `file`, returns an opaque origin otherwise, and never calls itself. A `blob:` URL with another inner scheme that previously produced a tuple origin, such as `blob:ws://...` or `blob:blob:https://...`, now returns an opaque origin.&lt;/p&gt;</content:encoded>
      <guid isPermaLink="false">https://vulnerability.circl.lu/vuln/rustsec-2026-0592</guid>
      <pubDate>Fri, 20 Nov 2026 12:00:00 +0000</pubDate>
    </item>
    <item>
      <title>RUSTSEC-2026-0591 — Ineffective default DACL on Windows</title>
      <link>https://vulnerability.circl.lu/vuln/rustsec-2026-0591</link>
      <description>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: secmem-proc&lt;/p&gt;
&lt;p&gt;On Windows, affected versions of this crate do not provide the intended security-enhancement.
In particular, any process running under the same user as the hardened process can reset our DACL,
and thus bypass the protection the DACL should have provided. Usage of the crate however *does not
actively harm security*.&lt;/p&gt;
&lt;p&gt;On affected version, an explicit owner SID (taken from the process token) was used for the DACL.
However, Windows implicitly grants extra access rights to the well-known &amp;#34;owner rights&amp;#34; SID, unless
it occurs in the DACL, which it did not. This allows any process under the same user to reset the
DACL of our process, making this defence effectively useless. The new code uses the well-known
&amp;#34;owner rights&amp;#34; SID, preventing the attack.&lt;/p&gt;
&lt;p&gt;The default hardening setup is fixed in version `0.3.9`. Users creating a custom DACL with the API
from the `win_acl` module need to manually update their code to get the security enhancement. They
are strongly advised to use a well-known &amp;#34;owner rights&amp;#34; SID created using the new
`win_acl::SidRef::well_known` (introduced in version `0.3.9`) in their DACL instead of the owner
SID returned by `win_acl::TokenUser::sid`.&lt;/p&gt;</description>
      <content:encoded>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: secmem-proc&lt;/p&gt;
&lt;p&gt;On Windows, affected versions of this crate do not provide the intended security-enhancement.
In particular, any process running under the same user as the hardened process can reset our DACL,
and thus bypass the protection the DACL should have provided. Usage of the crate however *does not
actively harm security*.&lt;/p&gt;
&lt;p&gt;On affected version, an explicit owner SID (taken from the process token) was used for the DACL.
However, Windows implicitly grants extra access rights to the well-known &amp;#34;owner rights&amp;#34; SID, unless
it occurs in the DACL, which it did not. This allows any process under the same user to reset the
DACL of our process, making this defence effectively useless. The new code uses the well-known
&amp;#34;owner rights&amp;#34; SID, preventing the attack.&lt;/p&gt;
&lt;p&gt;The default hardening setup is fixed in version `0.3.9`. Users creating a custom DACL with the API
from the `win_acl` module need to manually update their code to get the security enhancement. They
are strongly advised to use a well-known &amp;#34;owner rights&amp;#34; SID created using the new
`win_acl::SidRef::well_known` (introduced in version `0.3.9`) in their DACL instead of the owner
SID returned by `win_acl::TokenUser::sid`.&lt;/p&gt;</content:encoded>
      <guid isPermaLink="false">https://vulnerability.circl.lu/vuln/rustsec-2026-0591</guid>
      <pubDate>Tue, 27 Oct 2026 12:00:00 +0000</pubDate>
    </item>
    <item>
      <title>RUSTSEC-2026-0590 — `Fixed` and `Dynamic` secret wrappers were not zeroized on drop</title>
      <link>https://vulnerability.circl.lu/vuln/rustsec-2026-0590</link>
      <description>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: secure-gate&lt;/p&gt;
&lt;p&gt;In affected versions, the `Fixed&amp;lt;T&amp;gt;` and `Dynamic&amp;lt;T&amp;gt;` wrappers (and the
`FixedNoClone` / `DynamicNoClone` variants in 0.6.x) did not wipe their contents
when dropped. The types implemented the `zeroize::ZeroizeOnDrop` marker trait,
which has no methods and does nothing by itself, but had no `Drop` implementation.
The documentation stated that secrets were zeroized on drop.&lt;/p&gt;
&lt;p&gt;A secret held in one of these wrappers stayed in the freed stack or heap memory
after the wrapper went out of scope, where it could be read by anything able to
observe that memory later, such as a core dump, swap, or a separate memory
disclosure bug.&lt;/p&gt;
&lt;p&gt;Versions before 0.5.0 are not affected: with the default `zeroize` feature they
stored the secret in `secrecy::SecretBox` or `zeroize::Zeroizing`, both of which
wipe on drop.&lt;/p&gt;
&lt;p&gt;The flaw was fixed by making `zeroize` a required dependency, bounding the
wrappers on `T: Zeroize`, and adding `Drop` implementations that call
`zeroize()`. The first published release with the fix is 0.8.0-rc.1; the first
stable release is 0.8.0. All affected versions have been yanked from crates.io.&lt;/p&gt;</description>
      <content:encoded>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: secure-gate&lt;/p&gt;
&lt;p&gt;In affected versions, the `Fixed&amp;lt;T&amp;gt;` and `Dynamic&amp;lt;T&amp;gt;` wrappers (and the
`FixedNoClone` / `DynamicNoClone` variants in 0.6.x) did not wipe their contents
when dropped. The types implemented the `zeroize::ZeroizeOnDrop` marker trait,
which has no methods and does nothing by itself, but had no `Drop` implementation.
The documentation stated that secrets were zeroized on drop.&lt;/p&gt;
&lt;p&gt;A secret held in one of these wrappers stayed in the freed stack or heap memory
after the wrapper went out of scope, where it could be read by anything able to
observe that memory later, such as a core dump, swap, or a separate memory
disclosure bug.&lt;/p&gt;
&lt;p&gt;Versions before 0.5.0 are not affected: with the default `zeroize` feature they
stored the secret in `secrecy::SecretBox` or `zeroize::Zeroizing`, both of which
wipe on drop.&lt;/p&gt;
&lt;p&gt;The flaw was fixed by making `zeroize` a required dependency, bounding the
wrappers on `T: Zeroize`, and adding `Drop` implementations that call
`zeroize()`. The first published release with the fix is 0.8.0-rc.1; the first
stable release is 0.8.0. All affected versions have been yanked from crates.io.&lt;/p&gt;</content:encoded>
      <guid isPermaLink="false">https://vulnerability.circl.lu/vuln/rustsec-2026-0590</guid>
      <pubDate>Tue, 17 Mar 2026 12:00:00 +0000</pubDate>
    </item>
    <item>
      <title>RUSTSEC-2026-0589 — Validating Matrix key IDs may panic</title>
      <link>https://vulnerability.circl.lu/vuln/rustsec-2026-0589</link>
      <description>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: ruma-identifiers-validation&lt;/p&gt;
&lt;p&gt;Using `key_id::validate` may panic if the `algorithm` part of the key ID is
longer than 255 bytes and contains non-ASCII characters. The affected versions
assumed that the length of the first part containing the encryption or signing
algorithm (the part before the colon) would fit into a `NonZeroU8`, but the
length of the first part wasn&amp;#39;t validated.&lt;/p&gt;</description>
      <content:encoded>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: ruma-identifiers-validation&lt;/p&gt;
&lt;p&gt;Using `key_id::validate` may panic if the `algorithm` part of the key ID is
longer than 255 bytes and contains non-ASCII characters. The affected versions
assumed that the length of the first part containing the encryption or signing
algorithm (the part before the colon) would fit into a `NonZeroU8`, but the
length of the first part wasn&amp;#39;t validated.&lt;/p&gt;</content:encoded>
      <guid isPermaLink="false">https://vulnerability.circl.lu/vuln/rustsec-2026-0589</guid>
      <pubDate>Thu, 19 Nov 2026 12:00:00 +0000</pubDate>
    </item>
    <item>
      <title>RUSTSEC-2026-0588 — Validating Matrix Content URIs may panic</title>
      <link>https://vulnerability.circl.lu/vuln/rustsec-2026-0588</link>
      <description>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: ruma-identifiers-validation&lt;/p&gt;
&lt;p&gt;Using `mxc_uri::validate` may panic if the server name part of the URI is larger
than 249 bytes. The affected versions assumed that the length of the first part
of the URI (`mxc://{server_name}`) would fit into a `NonZeroU8`, but the length
of the server name wasn&amp;#39;t validated and the Matrix specification actually allows
the server name to have a maximum length of 261 bytes.&lt;/p&gt;</description>
      <content:encoded>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: ruma-identifiers-validation&lt;/p&gt;
&lt;p&gt;Using `mxc_uri::validate` may panic if the server name part of the URI is larger
than 249 bytes. The affected versions assumed that the length of the first part
of the URI (`mxc://{server_name}`) would fit into a `NonZeroU8`, but the length
of the server name wasn&amp;#39;t validated and the Matrix specification actually allows
the server name to have a maximum length of 261 bytes.&lt;/p&gt;</content:encoded>
      <guid isPermaLink="false">https://vulnerability.circl.lu/vuln/rustsec-2026-0588</guid>
      <pubDate>Thu, 19 Nov 2026 12:00:00 +0000</pubDate>
    </item>
    <item>
      <title>RUSTSEC-2026-0587 — WASI filesystem sandbox escape on Windows through a drive-prefixed path component</title>
      <link>https://vulnerability.circl.lu/vuln/rustsec-2026-0587</link>
      <description>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: wasmtime-wasi&lt;/p&gt;
&lt;p&gt;This is an entry in the RustSec database for the Wasmtime security advisory
located at
https://github.com/bytecodealliance/wasmtime/security/advisories/GHSA-f2pq-268w-886c
For more information see the GitHub-hosted security advisory.&lt;/p&gt;</description>
      <content:encoded>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: wasmtime-wasi&lt;/p&gt;
&lt;p&gt;This is an entry in the RustSec database for the Wasmtime security advisory
located at
https://github.com/bytecodealliance/wasmtime/security/advisories/GHSA-f2pq-268w-886c
For more information see the GitHub-hosted security advisory.&lt;/p&gt;</content:encoded>
      <guid isPermaLink="false">https://vulnerability.circl.lu/vuln/rustsec-2026-0587</guid>
      <pubDate>Thu, 19 Nov 2026 12:00:00 +0000</pubDate>
    </item>
    <item>
      <title>RUSTSEC-2026-0586 — ML-DSA verification accepts non-conforming signatures with the `avx2` feature</title>
      <link>https://vulnerability.circl.lu/vuln/rustsec-2026-0586</link>
      <description>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: pqcrypto-mldsa&lt;/p&gt;
&lt;p&gt;With the default `avx2` feature on x86_64 (non-Windows) and a CPU that supports
AVX2, `pqcrypto-mldsa` verifies signatures with PQClean&amp;#39;s AVX2 implementation of
ML-DSA. That implementation predates two upstream fixes
([pq-crystals/dilithium#113](https://github.com/pq-crystals/dilithium/issues/113),
3 June 2026):&lt;/p&gt;
&lt;p&gt;- the verifier does not check that the infinity norm of `z` is below
  `γ1 − β` (FIPS 204, Algorithm 8, line 13), so it accepts signatures that
  conforming verifiers reject;
- `use_hint_avx()` mishandles a boundary case, so one class of valid
  signatures is rejected.&lt;/p&gt;
&lt;p&gt;Creating an accepted out-of-bound signature requires the private key. The
practical effect is that verification results differ between this crate with
AVX2 and conforming implementations, including this crate without AVX2.&lt;/p&gt;
&lt;p&gt;The Wycheproof ML-DSA verification vectors reproduce both defects: with AVX2,
37 / 41 / 49 invalid vectors (ML-DSA-44 / -65 / -87) are accepted and one valid
vector per parameter set is rejected.&lt;/p&gt;
&lt;p&gt;The crate is unmaintained (see RUSTSEC-2026-0166) and no patched version is
expected. Disabling the `avx2` feature avoids the affected code. Migrating to a
maintained implementation, such as the [`ml-dsa`](https://crates.io/crates/ml-dsa)
crate, is recommended.&lt;/p&gt;</description>
      <content:encoded>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: pqcrypto-mldsa&lt;/p&gt;
&lt;p&gt;With the default `avx2` feature on x86_64 (non-Windows) and a CPU that supports
AVX2, `pqcrypto-mldsa` verifies signatures with PQClean&amp;#39;s AVX2 implementation of
ML-DSA. That implementation predates two upstream fixes
([pq-crystals/dilithium#113](https://github.com/pq-crystals/dilithium/issues/113),
3 June 2026):&lt;/p&gt;
&lt;p&gt;- the verifier does not check that the infinity norm of `z` is below
  `γ1 − β` (FIPS 204, Algorithm 8, line 13), so it accepts signatures that
  conforming verifiers reject;
- `use_hint_avx()` mishandles a boundary case, so one class of valid
  signatures is rejected.&lt;/p&gt;
&lt;p&gt;Creating an accepted out-of-bound signature requires the private key. The
practical effect is that verification results differ between this crate with
AVX2 and conforming implementations, including this crate without AVX2.&lt;/p&gt;
&lt;p&gt;The Wycheproof ML-DSA verification vectors reproduce both defects: with AVX2,
37 / 41 / 49 invalid vectors (ML-DSA-44 / -65 / -87) are accepted and one valid
vector per parameter set is rejected.&lt;/p&gt;
&lt;p&gt;The crate is unmaintained (see RUSTSEC-2026-0166) and no patched version is
expected. Disabling the `avx2` feature avoids the affected code. Migrating to a
maintained implementation, such as the [`ml-dsa`](https://crates.io/crates/ml-dsa)
crate, is recommended.&lt;/p&gt;</content:encoded>
      <guid isPermaLink="false">https://vulnerability.circl.lu/vuln/rustsec-2026-0586</guid>
      <pubDate>Fri, 20 Nov 2026 12:00:00 +0000</pubDate>
    </item>
    <item>
      <title>RUSTSEC-2026-0585 — ML-KEM does not perform the FIPS 203 input checks</title>
      <link>https://vulnerability.circl.lu/vuln/rustsec-2026-0585</link>
      <description>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: pqcrypto-mlkem&lt;/p&gt;
&lt;p&gt;`pqcrypto-mlkem` wraps PQClean&amp;#39;s ML-KEM implementation, which does not perform
the input checks required by FIPS 203:&lt;/p&gt;
&lt;p&gt;- encapsulation does not perform the encapsulation key modulus check
  (section 7.2), so encapsulation keys with coefficients that are not reduced
  modulo q are accepted;
- decapsulation does not perform the decapsulation key hash check
  (section 7.3), so a decapsulation key whose stored hash or embedded
  encapsulation key has been altered is accepted.&lt;/p&gt;
&lt;p&gt;With the Wycheproof ML-KEM vectors, 108 / 112 / 116 invalid encapsulation keys
(ML-KEM-512 / -768 / -1024) and 2 invalid decapsulation keys per parameter set
are accepted.&lt;/p&gt;
&lt;p&gt;This does not disclose key material. It is a FIPS 203 non-conformance, and an
affected implementation can disagree with a conforming one about which keys
are valid.&lt;/p&gt;
&lt;p&gt;The crate is unmaintained (see RUSTSEC-2026-0161) and no patched version is
expected. Migrating to a maintained implementation, such as the
[`ml-kem`](https://crates.io/crates/ml-kem) crate, is recommended.&lt;/p&gt;</description>
      <content:encoded>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: pqcrypto-mlkem&lt;/p&gt;
&lt;p&gt;`pqcrypto-mlkem` wraps PQClean&amp;#39;s ML-KEM implementation, which does not perform
the input checks required by FIPS 203:&lt;/p&gt;
&lt;p&gt;- encapsulation does not perform the encapsulation key modulus check
  (section 7.2), so encapsulation keys with coefficients that are not reduced
  modulo q are accepted;
- decapsulation does not perform the decapsulation key hash check
  (section 7.3), so a decapsulation key whose stored hash or embedded
  encapsulation key has been altered is accepted.&lt;/p&gt;
&lt;p&gt;With the Wycheproof ML-KEM vectors, 108 / 112 / 116 invalid encapsulation keys
(ML-KEM-512 / -768 / -1024) and 2 invalid decapsulation keys per parameter set
are accepted.&lt;/p&gt;
&lt;p&gt;This does not disclose key material. It is a FIPS 203 non-conformance, and an
affected implementation can disagree with a conforming one about which keys
are valid.&lt;/p&gt;
&lt;p&gt;The crate is unmaintained (see RUSTSEC-2026-0161) and no patched version is
expected. Migrating to a maintained implementation, such as the
[`ml-kem`](https://crates.io/crates/ml-kem) crate, is recommended.&lt;/p&gt;</content:encoded>
      <guid isPermaLink="false">https://vulnerability.circl.lu/vuln/rustsec-2026-0585</guid>
      <pubDate>Fri, 20 Nov 2026 12:00:00 +0000</pubDate>
    </item>
    <item>
      <title>RUSTSEC-2026-0584 — gRPC `IpKeyExtractor` keys on client-supplied forwarding headers</title>
      <link>https://vulnerability.circl.lu/vuln/rustsec-2026-0584</link>
      <description>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: tokio-rate-limit&lt;/p&gt;
&lt;p&gt;`tonic_middleware::IpKeyExtractor` is documented as rate limiting gRPC requests by client IP address,
but on a standard tonic server it keys on client-supplied `x-forwarded-for` and `x-real-ip`
headers. In 0.5.0–0.9.1, the extractor reads those headers directly; in 0.10.0–0.11.0, it first
checks for a bare `std::net::SocketAddr` request extension—which `tonic::transport::Server` never
supplies, as it provides `TcpConnectInfo` instead—and falls back to the same headers. Direct
clients can therefore choose their rate-limit key to bypass per-IP rate limits and increase the
number of tracked keys; without forwarding headers, 0.5.0–0.9.1 allow requests without limiting
them and 0.10.0–0.11.0 reject them by default unless `.fail_open()` is enabled.&lt;/p&gt;
&lt;p&gt;Applications are affected if they use `GrpcRateLimitLayer` with `IpKeyExtractor` (the
`tonic-support` feature). The HTTP middleware&amp;#39;s `IpKeyExtractor` reads `ConnectInfo&amp;lt;SocketAddr&amp;gt;` and
is not affected.&lt;/p&gt;
&lt;p&gt;Version 0.11.1 fixes `IpKeyExtractor` to honor a server-supplied `SocketAddr` override first and then
read the TCP peer address from tonic&amp;#39;s `TcpConnectInfo` (which `tonic::transport::Server` attaches
to plaintext connections and connections using its built-in TLS), without ever reading forwarding
headers. Deployments behind a trusted proxy can use the new opt-in `ForwardedIpKeyExtractor`, which
reads a single configured header (`x-forwarded-for` or `x-real-ip`) and is only correct when that
proxy sets the header as documented.&lt;/p&gt;
&lt;p&gt;If yo…&lt;/p&gt;</description>
      <content:encoded>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: tokio-rate-limit&lt;/p&gt;
&lt;p&gt;`tonic_middleware::IpKeyExtractor` is documented as rate limiting gRPC requests by client IP address,
but on a standard tonic server it keys on client-supplied `x-forwarded-for` and `x-real-ip`
headers. In 0.5.0–0.9.1, the extractor reads those headers directly; in 0.10.0–0.11.0, it first
checks for a bare `std::net::SocketAddr` request extension—which `tonic::transport::Server` never
supplies, as it provides `TcpConnectInfo` instead—and falls back to the same headers. Direct
clients can therefore choose their rate-limit key to bypass per-IP rate limits and increase the
number of tracked keys; without forwarding headers, 0.5.0–0.9.1 allow requests without limiting
them and 0.10.0–0.11.0 reject them by default unless `.fail_open()` is enabled.&lt;/p&gt;
&lt;p&gt;Applications are affected if they use `GrpcRateLimitLayer` with `IpKeyExtractor` (the
`tonic-support` feature). The HTTP middleware&amp;#39;s `IpKeyExtractor` reads `ConnectInfo&amp;lt;SocketAddr&amp;gt;` and
is not affected.&lt;/p&gt;
&lt;p&gt;Version 0.11.1 fixes `IpKeyExtractor` to honor a server-supplied `SocketAddr` override first and then
read the TCP peer address from tonic&amp;#39;s `TcpConnectInfo` (which `tonic::transport::Server` attaches
to plaintext connections and connections using its built-in TLS), without ever reading forwarding
headers. Deployments behind a trusted proxy can use the new opt-in `ForwardedIpKeyExtractor`, which
reads a single configured header (`x-forwarded-for` or `x-real-ip`) and is only correct when that
proxy sets the header as documented.&lt;/p&gt;
&lt;p&gt;If yo…&lt;/p&gt;</content:encoded>
      <guid isPermaLink="false">https://vulnerability.circl.lu/vuln/rustsec-2026-0584</guid>
      <pubDate>Thu, 19 Nov 2026 12:00:00 +0000</pubDate>
    </item>
    <item>
      <title>RUSTSEC-2026-0583 — Panics during RSA private-key construction, validation, or re-encoding</title>
      <link>https://vulnerability.circl.lu/vuln/rustsec-2026-0583</link>
      <description>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: sad-rsa&lt;/p&gt;
&lt;p&gt;`RsaPrivateKey` private-key construction, validation, and serialization can panic across three
classes of inputs, as well as when re-encoding certain valid keys:&lt;/p&gt;
&lt;p&gt;- `RsaPrivateKey::from_components` and PKCS#1 or PKCS#8 private-key decoding (DER or PEM) can panic
  during validation when a supplied factor equals 1, because `RsaPrivateKey::validate()` let a
  factor of 1 reach a later check that panicked.
- `RsaPrivateKey::from_primes` and `RsaPrivateKey::from_p_q` can panic during construction when a
  supplied factor is 0, 1, or even.
- Prime recovery in `RsaPrivateKey::from_components` can panic when the private exponent is 0.
- For a key, including a valid one, whose first prime occupies more limbs than the second,
  `crt_coefficient`, `to_pkcs1_der`, `to_pkcs8_der`, and `serde` serialization can panic.&lt;/p&gt;
&lt;p&gt;An application that imports untrusted private keys or factors, or re-encodes such keys, can lose a
worker, or the whole process when built with `panic = &amp;#34;abort&amp;#34;`. Public-key parsing is not affected.&lt;/p&gt;
&lt;p&gt;Fixed in 0.10.4, which rejects factors below 2, even factors, repeated factors, and a zero private
exponent during prime recovery, and computes the CRT coefficient and re-encodes valid two-prime
keys when the first prime occupies more limbs than the second.&lt;/p&gt;
&lt;p&gt;If you cannot upgrade to 0.10.4, isolate the entire import, construction, CRT calculation, or
serialization in a separate process and treat an abnormal exit as an operation failure, or use
`catch_unwind` around the entire…&lt;/p&gt;</description>
      <content:encoded>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; crates.io: sad-rsa&lt;/p&gt;
&lt;p&gt;`RsaPrivateKey` private-key construction, validation, and serialization can panic across three
classes of inputs, as well as when re-encoding certain valid keys:&lt;/p&gt;
&lt;p&gt;- `RsaPrivateKey::from_components` and PKCS#1 or PKCS#8 private-key decoding (DER or PEM) can panic
  during validation when a supplied factor equals 1, because `RsaPrivateKey::validate()` let a
  factor of 1 reach a later check that panicked.
- `RsaPrivateKey::from_primes` and `RsaPrivateKey::from_p_q` can panic during construction when a
  supplied factor is 0, 1, or even.
- Prime recovery in `RsaPrivateKey::from_components` can panic when the private exponent is 0.
- For a key, including a valid one, whose first prime occupies more limbs than the second,
  `crt_coefficient`, `to_pkcs1_der`, `to_pkcs8_der`, and `serde` serialization can panic.&lt;/p&gt;
&lt;p&gt;An application that imports untrusted private keys or factors, or re-encodes such keys, can lose a
worker, or the whole process when built with `panic = &amp;#34;abort&amp;#34;`. Public-key parsing is not affected.&lt;/p&gt;
&lt;p&gt;Fixed in 0.10.4, which rejects factors below 2, even factors, repeated factors, and a zero private
exponent during prime recovery, and computes the CRT coefficient and re-encodes valid two-prime
keys when the first prime occupies more limbs than the second.&lt;/p&gt;
&lt;p&gt;If you cannot upgrade to 0.10.4, isolate the entire import, construction, CRT calculation, or
serialization in a separate process and treat an abnormal exit as an operation failure, or use
`catch_unwind` around the entire…&lt;/p&gt;</content:encoded>
      <guid isPermaLink="false">https://vulnerability.circl.lu/vuln/rustsec-2026-0583</guid>
      <pubDate>Thu, 19 Nov 2026 12:00:00 +0000</pubDate>
    </item>
  </channel>
</rss>
