tavily-mcp

Pass

Audited by Gen Agent Trust Hub on Feb 21, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [Indirect Prompt Injection] (LOW): The skill is designed to fetch content from external, untrusted web sources through tools like tavily_extract, tavily_search, and tavily_crawl.
  • Ingestion points: Web page content and search snippets enter the agent's context through tavily_extract, tavily_search, tavily_crawl, and tavily_research (SKILL.md).
  • Boundary markers: None identified in the skill instructions to separate external content from internal logic.
  • Capability inventory: The skill allows the agent to perform extensive web operations including multi-page crawling and deep research across multiple sources (SKILL.md).
  • Sanitization: No explicit sanitization or instruction-ignoring patterns are provided for handling malicious instructions embedded in fetched web content.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 21, 2026, 02:26 AM