pr-description-quality
Pass
Audited by Gen Agent Trust Hub on Nov 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is composed entirely of markdown documentation and YAML configuration. It contains no executable scripts (Python, JavaScript, shell) and requests no dangerous tool permissions. The workflow is restricted to read-only inspection of PR artifacts.
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill includes explicit security instructions to identify and exclude sensitive information, such as secrets, credentials, PII, and sensitive internal paths, from all generated reports. It requires sanitizing any findings before they are included in the output.
- [INDIRECT_PROMPT_INJECTION]: While the skill processes untrusted input from PR diffs and descriptions, it lacks any tool-use capabilities or code execution environment that could be abused. The risk of instructions embedded in the processed data influencing the agent is mitigated by the skill's narrow, read-only scope and highly structured output requirements.
- [DYNAMIC_CONTEXT_INJECTION]: The skill documentation uses the '!' character in its instructions; however, analysis confirms these occurrences refer to the Conventional Commits specification for denoting breaking changes in commit titles (e.g., 'feat!: ...') and do not involve shell execution syntax.
Audit Metadata