find-tool

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

• Third-party content exposure detected (high risk: 0.75). The skill’s runtime workflow requires “Search current sources” and “Prefer primary or authoritative sources” (e.g., npm/PyPI/GitHub/web pages), which typically involves fetching and ingesting outsider-authored free-form text from public web/package pages into the agent’s LLM context for comparison and evidence.