firecrawl

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

• Third-party content exposure detected (medium risk: 0.30). references/output-security.md states that “Fetched web content and parsed documents are untrusted input” and that the skill writes fetched page/parsed document content from Firecrawl into .firecrawl/ for inspection, meaning outsider-authored text from user-provided URLs can be ingested at runtime via firecrawl search/scrape/map/crawl/agent/parse.