Paper/Archive/CVE-2025-1550

CVE-2025-1550

Google Keras — Arbitrary Code Execution via Crafted Keras Config for Model Loading

Description · from the CVE record

The Keras Model.load_model function permits arbitrary code execution, even with safe_mode=True, through a manually constructed, malicious .keras archive. By altering the config.json file within the archive, an attacker can specify arbitrary Python modules and functions, along with their arguments, to be loaded and executed during model loading.

Affected

ProductAffectedFixed inSource
Keras3.0.0 – < 3.8.03.8.0CNA

Scores by source

NVD9.8CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
Google (CNA)7.3CVSS:4.0/AV:L/AC:L/AT:P/PR:L/UI:A/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H
CISA ADPSSVCExploitation: poc · Automatable: no · Technical impact: total

References

refgithub.com/keras-team/keras/pull/20751reftowerofhanoi.it/writeups/cve-2025-1550