AIKIDO-2025-10413
openai-whisper is vulnerable to Unsafe Deserialization
65
Medium Risk
This Affects:
openai-whisperTL;DR
Affected versions of this package load models using torch.load() without setting the weights_only parameter to True, thereby enabling full pickle deserialization. Since pickle is inherently insecure, this allows the execution of arbitrary code during model loading. A remote attacker could exploit this behavior by crafting malicious pickle data, potentially leading to serious vulnerabilities such as remote code execution (RCE).
Who does this affect?
You are affected if you are using a version that falls within the vulnerable range.
Background info
openai-whisper is vulnerable to Unsafe Deserialization in versions 20230117 - 20240930.
How to fix this
Upgrade the openai-whisper library to the patch version.
Are You Affected?
Connect your repositories to instantly see whether vulnerable or malicious packages exist in your codebase.
Free. No credit card required.
SOC 2Compliant
ISO 27001Compliant