import-model

Warn

Audited by Snyk on Jun 13, 2026

Risk Level: MEDIUM
Full Analysis

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

  • Potentially malicious external URL detected (high risk: 1.00). The repo's scripts fetch Hugging Face model repos at runtime (e.g. https://huggingface.co/<HF_MODEL_ID>/raw/main/config.json and related safetensors/model files) and call transformers.from_pretrained(..., trust_remote_code=True) which downloads and can execute remote model code from the HF repo, meaning external content is fetched at runtime, may execute code that controls model behavior, and is required for the skill.

Issues (1)

W012
MEDIUM

Unverifiable external dependency detected (runtime URL that controls agent).

Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 13, 2026, 05:20 PM
Issues
1
Security Audit — snyk — import-model