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.
Metrics
Affected Vendors & Products
No data.
No data.
No data.
No data.
References
History
Mon, 13 Apr 2026 15:00:00 +0000
| Type | Values Removed | Values Added |
|---|---|---|
| Description | 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. | |
| Title | Safe Mode Bypass in keras-team/keras | |
| Weaknesses | CWE-502 | |
| References |
| |
| Metrics |
cvssV3_0
|
MITRE
Status: PUBLISHED
Assigner: @huntr_ai
Published:
Updated: 2026-04-13T14:55:28.649Z
Reserved: 2026-01-27T04:14:51.848Z
Link: CVE-2026-1462
Vulnrichment
No data.
NVD
Status : Received
Published: 2026-04-13T15:17:18.967
Modified: 2026-04-13T15:17:18.967
Link: CVE-2026-1462
Redhat
No data.
OpenCVE Enrichment
No data.