address-pr-review
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from GitHub PR comments, creating a potential attack surface for indirect prompt injection.\n
- Ingestion points: Comment bodies are fetched via
gh api graphqlinSKILL.md.\n - Boundary markers: The instructions include a comprehensive "Critical evaluation" section requiring the agent to verify comments against PR descriptions, ticket criteria, and the current codebase state before acting.\n
- Capability inventory: The skill can modify local files, execute build/test commands, and use
ghto post replies or push code to GitHub.\n - Sanitization: The skill relies on the agent's skepticism and the requirement for a "minimal diff" to mitigate the risk of malicious suggestions being implemented.\n- [EXTERNAL_DOWNLOADS]: The skill fetches pull request metadata and comments from GitHub.\n
- Evidence: Uses
gh apito interact with GitHub repositories. GitHub is a well-known and trusted platform for software development, and the operations are restricted to the active PR context.
Audit Metadata