requesting-code-review
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes local shell commands to interact with the git repository and run tests.
- Evidence: Usage of
git rev-parse,git log, andrtk npm testwithinSKILL.mdto establish review context and verify work. - [DYNAMIC_EXECUTION]: The skill generates subagent prompts by interpolating implementation details and git commit identifiers into a local markdown template.
- Evidence: Template placeholder substitution in
code-reviewer.mdused for subagent dispatch. - [INDIRECT_PROMPT_INJECTION]: The review subagent ingests code diffs which could contain instructions intended to influence agent behavior. The skill mitigates risks by including specific protocols to redact secrets and PII.
- Ingestion points: Git diff output processed by the subagent defined in
code-reviewer.md. - Boundary markers: Structured report format in the subagent template helps maintain focus on code analysis.
- Capability inventory: The subagent is primarily tasked with report generation and quality assessment.
- Sanitization: Explicit instructions to redact sensitive data and PII before sharing context with subagents.
Audit Metadata