Skip to content
Threat Feed

Tag

Huggingface

1 brief RSS
critical advisory

InstructLab Arbitrary Code Execution via Malicious HuggingFace Model

InstructLab is vulnerable to arbitrary code execution because the `linux_train.py` script hardcodes `trust_remote_code=True` when loading models from HuggingFace, allowing remote attackers to execute code by convincing a user to load a malicious model.

InstructLab cve code-execution huggingface
2r 1t 1c