exa-ci-integration
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill explicitly instructs against hardcoding or exposing credentials, recommending the injection of
EXA_API_KEYfrom approved server-side secret managers. - [SAFE]: The instructions mandate the use of sanitized fixtures and the redaction of sensitive content (retrieved data, raw queries, generated output) from CI evidence.
- [SAFE]: External references target official documentation at
exa.ai, which is a well-known and established service. - [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing external vendor data (Exa API responses) within CI pipelines.
- Ingestion points: Processes request/response schemas and live search results within CI workflows as described in
SKILL.md. - Boundary markers: The skill explicitly instructs the use of "sanitized fixtures" and "content-free gate evidence" to delimit raw external data from the agent context.
- Capability inventory: The skill is restricted to file operations (
Read,Glob,Grep,Write,Edit) and does not utilize shell execution or direct network tools for the agent itself. - Sanitization: Instructions require verifying that sensitive fields are absent from evidence and excluding raw retrieved content unless separately approved.
Audit Metadata