tw-research-proposal-diamond

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

• Third-party content exposure detected (high risk: 0.90). The skill's workflow (SKILL.md phases 1–4, e.g., "立刻使用 web_search 搜尋" in Phase 1 and "使用 web_search 補充具體引用文獻（至少 5 筆）" in Phase 4) explicitly instructs the agent to fetch and ingest open/public web content and use those results to shape research directions and outputs, so untrusted third‑party content can materially influence its actions.