hugging-face-cli

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.75). The required runtime workflow for this skill is the Hugging Face CLI’s Hub browsing/downloading commands (e.g., hf models ls/info, hf datasets ls/info, hf spaces ls/info, hf download), which fetches outsider-authored repository metadata and files from public Hugging Face Hub pages/APIs and can ingest their free-text fields (e.g., model/dataset/space descriptions/cardData) into the agent’s LLM context via command outputs.