research

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow uses autonomous research agents that perform web/URL scraping and search (e.g., Perplexity/Exa/Firecrawl “scraping specific URLs” and “browser-extracted content”), so outsider-authored free text from public web pages can be ingested into the LLM context via those MCP tool calls.