receiving-regression-review
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data such as pull request comments, regression-review reports, and unstructured feedback to identify and fix software regressions.
- Ingestion points: External reports and feedback processed in SKILL.md workflow step 1.
- Boundary markers: The instructions do not define specific delimiters or "ignore" instructions for the ingested content.
- Capability inventory: The skill is authorized to perform code modifications in the repository (workflow step 6).
- Sanitization: There is no explicit requirement to sanitize or escape the content of the external reports before processing.
Audit Metadata