expert-research

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly performs 3-angle web searches (Step 2.2: mcp__tavily__tavily_search and Exa), scrapes and summarizes top sources (Step 2.3), and uses Perplexity + a critic sub-agent to read/interpret those public sources (including HN, vendor blogs, community posts) to drive fact-checks, recommendations, and decision-making, so untrusted third‑party content can materially influence agent actions.