tw-research-hypothesis-generator

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

• Third-party content exposure detected (high risk: 0.90). Yes — the skill's required Step 3 ("文獻確認搜尋") explicitly performs 2–3 WebSearches of public sources (e.g., site:scholar and web queries) and instructs the agent to summarize and use those search results to generate and rank competing hypotheses and design methods, meaning untrusted third‑party content from the open web is read, interpreted, and can materially change the agent's outputs.