deep-research

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

• Third-party content exposure detected (high risk: 0.90). The skill's SKILL.md Research Pipeline (Phase 2: "Search" and Phase 3: "Read") and the Tool Discovery table explicitly require using web search/fetch tools (e.g., WebSearch, web_fetch, firecrawl_search, firecrawl_scrape, read_url) to retrieve and read full pages from the open web — and the quality-standards explicitly allow and annotate Tier‑5 sources (forums/social media/general websites) — meaning the agent will ingest untrusted third-party content that can materially influence synthesis and recommendations.