hugging-face-evaluation

Pass

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • API Token Management: The skill requires HF_TOKEN and AA_API_KEY for its operations. Documentation correctly instructs users to manage these via environment variables or .env files rather than hardcoding them, following standard security practices.
  • External Data Ingestion: The workflow fetches content from Hugging Face model cards (READMEs), the Artificial Analysis API, and academic papers via arXiv. This represents a potential surface for indirect prompt injection, as the agent processes untrusted text from external authors to extract scores. The implementation uses regular expressions and normalization to mitigate parsing errors.
  • Command Line Operations: The skill utilizes the Hugging Face CLI (hf) and uv for script execution and PR management. These operations are conducted within the /tmp/ directory and use standard authentication mechanisms.
  • Automated Script Execution: Several Python scripts are included to automate data mapping and PR creation. These scripts use well-known dependencies (like requests, pyyaml, and huggingface-hub) and are intended to be run in a local environment by the user or an authorized agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 18, 2026, 11:10 AM
Security Audit — agent-trust-hub — hugging-face-evaluation