deep-research-generic
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external, untrusted sources (the public internet), creating a surface for indirect prompt injection attacks.
- Ingestion points: Information retrieved via search engines, web browsers (Playwright), site crawlers, and the Tavily research API, as described in SKILL.md.
- Boundary markers: The skill does not implement specific text delimiters within its instructions but explicitly directs the agent to treat source content as untrusted.
- Capability inventory: The skill utilizes network-enabled tools (Search, Browser, Tavily) and performs file system operations (writing research records to the
.research/directory). - Sanitization: The instructions recommend loading an external hygiene skill (
untrusted-input-hygiene) when processing instructions found within external sources.
Audit Metadata