Keras has an untrusted deserialization vulnerability
A vulnerability in the `TFSMLayer` class of the `keras` package, version 3.13.0, allows attacker-controlled TensorFlow SavedModels to be loaded during deserialization of `.keras` models, even when `safe_mode=True`. This bypasses the security guarantees of `safe_mode` and enables arbitrary attacker-controlled code execution during model inference under the victim's privileges. The issue arises due to the unconditional loading of external SavedModels, serialization of attacker-controlled file paths, and the lack of validation in the `from_config()` method.
describes (1)
Related CVEs.
| ID | severity | exploitable | ||
|---|---|---|---|---|
| CVE-2026-1462US | — | — | ||
A vulnerability in the `TFSMLayer` class of the `keras` package, version 3.13.0, allows attacker-controlled TensorFlow SavedModels to be loaded during deserialization of `.keras` m… | ||||
references
External citations — patches, advisories, analysis.
- https://github.com/advisories/GHSA-4f3f-g24h-fr8mADVISORY
- https://github.com/keras-team/kerasPACKAGE
- https://github.com/keras-team/keras/commit/b6773d3decaef1b05d8e794458e148cb362f163fWEB
- https://github.com/keras-team/keras/pull/22035WEB
- https://huntr.com/bounties/7e78d6f1-6977-4300-b595-e81bdbda331cWEB
- https://nvd.nist.gov/vuln/detail/CVE-2026-1462ADVISORY
- https://pypi.org/project/kerasPACKAGE
Advisories like this name the exposed products attackers go after on your perimeter. See what is exposed on your perimeter with a free passive scan.
Free scan