code-review
Pass
Audited by Gen Agent Trust Hub on Mar 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local shell commands to manage the review workflow. Evidence: git diff is used to check for local changes, git symbolic-ref identifies the default branch, and coderabbit review triggers the analysis.
- [EXTERNAL_DOWNLOADS]: The skill interacts with an external service via the CodeRabbit CLI. Evidence: Codebase diffs are sent to CodeRabbit for automated review as part of the primary skill function.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface when processing external tool output.
- Ingestion points: Output from the coderabbit review command (SKILL.md).
- Boundary markers: Looks for a specific Review completed completion marker.
- Capability inventory: Shell command execution for Git and CodeRabbit.
- Sanitization: Employs a fixed Markdown schema and explicitly forbids the agent from adding manual findings or subjective triage.
Audit Metadata