security
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted source code using 'Read', 'Glob', and 'Grep' tools. This creates an attack surface where specially crafted code comments or documentation within the analyzed project could attempt to influence the agent's output. However, this is the primary intended function of the skill, and the risk is mitigated by the agent's focus on structured security reporting.
- Ingestion points: File reading and grepping of local project source code in Phase 1 and Quick Checks.
- Boundary markers: None explicitly defined for the data ingestion phase.
- Capability inventory: Filesystem read access (Read, Glob, Grep).
- Sanitization: Not applicable as the skill is for static code analysis.
- [EXTERNAL_DOWNLOADS]: The skill references a well-known third-party security library, TrustKit, via its official GitHub repository (github.com/datatheorem/TrustKit.git). This is provided as a recommendation for production-grade certificate pinning and does not involve automated or hidden execution of untrusted code.
Audit Metadata