hugging-science

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime path uses scripts/fetch_catalog.py to fetch and parse outsider-controlled free text from https://huggingscience.co/topics/<slug>.md / llms-full.txt / llms.txt, including arbitrary markdown descriptions and metadata that are ingested and then rendered for the LLM to consume.