tavily-deep-research
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill performs web searches and content extraction through the Tavily API via a SandBase gateway. This involves fetching data from various external URLs and websites to fulfill research requests.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process content from external websites via the
tavily_extractcapability. This represents an attack surface where instructions embedded in web content could attempt to influence the agent's subsequent behavior. - Ingestion points: Content extracted from URLs provided by users or search results.
- Boundary markers: None specified in the instructions.
- Capability inventory: Web search, content extraction, and site mapping via the SandBase tools.
- Sanitization: No explicit sanitization or filtering of extracted content is mentioned in the provided skill instructions.
Audit Metadata