exa-search
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The installation section provides a command (
npx skills add -g BenedictKing/exa-search) that fetches the skill's logic and configuration from a community GitHub repository not managed by a trusted vendor. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from search results, which is a common vector for indirect prompt injection attacks.
- Ingestion points: Semantic search snippets and structured research data retrieved via the Exa API as specified in the Step-by-Step Guide.
- Boundary markers: The skill body includes instructional warnings (e.g., "Do not claim returned search snippets establish facts by themselves") which serve as a behavioral boundary, though they do not programmatically sanitize the input.
- Capability inventory: The skill possesses the capability to perform web-based semantic searches and content retrieval.
- Sanitization: The instructions recommend verifying material claims at the linked source and reporting uncertainties, but no automated sanitization or filtering of the external content is described.
Audit Metadata