ito-diagnose

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

• Third-party content exposure detected (high risk: 0.80). The skill's mandatory Step 1.5 (外部 library 文件驗證) instructs the agent to fetch and read external documentation using tools like "find-docs", "deepwiki MCP", and "exa MCP (web 搜尋備援)" — i.e., public web/GitHub/docs — and those findings are required to influence diagnostic hypotheses and next actions, exposing the agent to untrusted third‑party content.