deep-research

Pass

Audited by Gen Agent Trust Hub on Oct 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill inherently processes untrusted third-party data from the web, which constitutes an indirect prompt injection surface.
  • Ingestion points: Web content retrieved via firecrawl_scrape and crawling_exa tools.
  • Boundary markers: None specified for separating scraped content from internal processing logic.
  • Capability inventory: Web search, web scraping, and file system write access.
  • Sanitization: The instructions do not specify sanitization for the scraped web content, though quality rules (Step 5/Quality Rules) encourage checking for facts vs inference.
  • [COMMAND_EXECUTION]: The skill allows for local file system interaction as part of its delivery mechanism.
  • Evidence: The 'Deliver' section instructions state to "save full report to a file" for long reports.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 14, 2026, 02:57 PM