research-briefing

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

• Third-party content exposure detected (high risk: 0.90). This skill's run.py calls tavily_search to fetch search results (including URLs and page content) from external web sources and then instructs the agent to generate a briefing "based on the above search results," which clearly ingests untrusted third‑party content that could carry indirect prompt injections.