Home » Database

AVID-2026-R0034

Description

Arbitrary Code Execution via Crafted Keras Config for Model Loading (CVE-2025-1550)

Details

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.

References

Affected or Relevant Artifacts

  • Developer: Google
  • Deployer: Google
  • Artifact Details:
TypeName
SystemKeras

Impact

AVID Taxonomy Categorization

  • Risk domains: Security
  • SEP subcategories: S0100: Software Vulnerability
  • Lifecycle stages: L06: Deployment

CWE

IDDescription
CWE-94CWE-94: Improper Control of Generation of Code (‘Code Injection’)

Other information

  • Report Type: Advisory
  • Credits:
  • Date Reported: 2025-03-11
  • Version: 0.3.1
  • AVID Entry