GHSA-5GWJ-M78Q-7PQ3
Vulnerability from github – Published: 2026-07-03 21:31 – Updated: 2026-08-07 20:30A vulnerability in keras-team/keras version 3.14.0 allows for arbitrary code execution due to improper handling of deserialization in the Lambda layer. Specifically, the _raise_for_lambda_deserialization() function fails to enforce the safe-mode guard when safe_mode is set to None, which is the default value when from_config() is called outside of a SafeModeScope context. This logic error conflates None (unset/default-deny) with False (explicitly disabled), bypassing the guard and allowing attacker-controlled marshal bytecode to be deserialized. Affected call sites include keras.layers.deserialize(config), keras.models.clone_model(model), and any direct invocation of Lambda.from_config(config) without an enclosing SafeModeScope(True). This vulnerability can be exploited to achieve arbitrary OS-level code execution in the context of the server or user process.
{
"affected": [
{
"package": {
"ecosystem": "PyPI",
"name": "keras"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "3.12.3"
}
],
"type": "ECOSYSTEM"
}
]
},
{
"package": {
"ecosystem": "PyPI",
"name": "keras"
},
"ranges": [
{
"events": [
{
"introduced": "3.13.0"
},
{
"fixed": "3.15.0"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-12481"
],
"database_specific": {
"cwe_ids": [
"CWE-502"
],
"github_reviewed": true,
"github_reviewed_at": "2026-08-07T20:30:41Z",
"nvd_published_at": "2026-07-03T21:16:54Z",
"severity": "HIGH"
},
"details": "A vulnerability in keras-team/keras version 3.14.0 allows for arbitrary code execution due to improper handling of deserialization in the `Lambda` layer. Specifically, the `_raise_for_lambda_deserialization()` function fails to enforce the safe-mode guard when `safe_mode` is set to `None`, which is the default value when `from_config()` is called outside of a `SafeModeScope` context. This logic error conflates `None` (unset/default-deny) with `False` (explicitly disabled), bypassing the guard and allowing attacker-controlled `marshal` bytecode to be deserialized. Affected call sites include `keras.layers.deserialize(config)`, `keras.models.clone_model(model)`, and any direct invocation of `Lambda.from_config(config)` without an enclosing `SafeModeScope(True)`. This vulnerability can be exploited to achieve arbitrary OS-level code execution in the context of the server or user process.",
"id": "GHSA-5gwj-m78q-7pq3",
"modified": "2026-08-07T20:30:41Z",
"published": "2026-07-03T21:31:36Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-12481"
},
{
"type": "WEB",
"url": "https://github.com/keras-team/keras/pull/23048"
},
{
"type": "WEB",
"url": "https://github.com/keras-team/keras/pull/23165"
},
{
"type": "WEB",
"url": "https://github.com/keras-team/keras/commit/55888d3becbdbb45dc16a55b489900f911e2dde5"
},
{
"type": "WEB",
"url": "https://github.com/keras-team/keras/commit/d338a45204bdc787c8b3c4a9b82c1911cd52dedf"
},
{
"type": "PACKAGE",
"url": "https://github.com/keras-team/keras"
},
{
"type": "WEB",
"url": "https://github.com/keras-team/keras/releases/tag/v3.12.3"
},
{
"type": "WEB",
"url": "https://github.com/keras-team/keras/releases/tag/v3.15.0"
},
{
"type": "WEB",
"url": "https://huntr.com/bounties/59ceaed1-c8a3-4135-8f94-169ade02823d"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.0/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"type": "CVSS_V3"
}
],
"summary": "Keras: Lambda deserialization can bypass safe mode and execute code"
}
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.