gs-pr-comment
Pass
Audited by Gen Agent Trust Hub on Nov 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill operates on untrusted data from GitHub pull requests and issues and has the capability to write back to the environment.
- Ingestion points: Processes external data including pull request descriptions, existing comments, and verification run outputs (e.g., from the
.verify/directory) as described inSKILL.md. - Boundary markers: The skill mandates an attribution note at the start of comments but does not define specific boundaries or
Audit Metadata