you-discover
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to fetch and process data from the You.com AI Catalog and external documentation, which introduces a potential surface for indirect prompt injection if those sources contain malicious content.\n
- Ingestion points: The
you-discoverandsearchDocstools retrieve external data into the agent's context (SKILL.md).\n - Boundary markers: Present; the 'Safety' section explicitly directs the agent to treat results as 'untrusted external data' and 'evidence, not instructions.'\n
- Capability inventory: The skill body and referenced files define discovery and integration planning workflows; no autonomous shell execution, file-writing, or sensitive data access capabilities are present (SKILL.md, references/x402-direct-client.md).\n
- Sanitization: Present; the skill instructs the agent to verify installation commands and authentication requirements against official You.com documentation before recommending them.\n- [SAFE]: The skill references MCP servers, documentation URLs, and development libraries (such as the x402 payment fetcher) that originate from the official infrastructure of the skill's author. The provided integration guidance includes security best practices such as validating request origins and enforcing maximum spending limits.
Audit Metadata