huggingface-datasets

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

• Third-party content exposure detected (high risk: 0.90). This skill performs read-only Dataset Viewer API calls to public Hugging Face dataset endpoints (e.g., /first-rows, /rows, /search, /parquet on https://datasets-server.huggingface.co) which return user-uploaded, untrusted dataset content that the agent is expected to read and use to drive searches, pagination, and follow-on queries.