Description

The Keras Model.load_model method can be exploited to achieve arbitrary code execution, even with safe_mode=True. One can create a specially crafted .h5/.hdf5 model archive that, when loaded via Model.load_model, will trigger arbitrary code to be executed. This is achieved by crafting a special .h5 archive file that uses the Lambda layer feature of keras which allows arbitrary Python code in the form of pickled code. The vulnerability comes from the fact that the safe_mode=True option is not honored when reading .h5 archives. Note that the .h5/.hdf5 format is a legacy format supported by Keras 3 for backwards compatibility.

INFO

Published Date :

2025-09-19T08:16:44.772Z

Last Modified :

2026-02-26T17:48:23.327Z

Source :

Google
AFFECTED PRODUCTS

The following products are affected by CVE-2025-9905 vulnerability.

Vendors Products
Keras
  • Keras
REFERENCES

Here, you will find a curated list of external links that provide in-depth information to CVE-2025-9905.

URL Resource
https://github.com/keras-team/keras/pull/21602 cve-icon cve-icon cve-icon
https://github.com/keras-team/keras/security/advisories/GHSA-36rr-ww3j-vrjv cve-icon cve-icon cve-icon
https://nvd.nist.gov/vuln/detail/CVE-2025-9905 cve-icon
https://www.cve.org/CVERecord?id=CVE-2025-9905 cve-icon
Login to claim this CVE

CVSS Vulnerability Scoring System

Detailed values of each vector for above chart.
Attack Vector
Attack Complexity
Attack Requirements
Privileges Required
User Interaction
VS Confidentiality
VS Integrity
VS Availability
SS Confidentiality
SS Integrity
SS Availability
Detailed values of each vector for above chart.
Attack Vector
Attack Complexity
Privileges Required
User Interaction
Scope
Confidentiality Impact
Integrity Impact
Availability Impact