review-comments
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill identifies the target files for review by executing local git commands, specifically git diff and git symbolic-ref. This is standard functionality for a git-integrated developer tool.- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted source code and comments, which could potentially contain instructions designed to manipulate the agent's review.
- Ingestion points: The agent reads source code files and git diff outputs as specified in the 대상 결정 step of SKILL.md.
- Boundary markers: No specific delimiters or boundary markers are used to isolate analyzed code from the skill's instructions.
- Capability inventory: The skill's capabilities are restricted to generating textual analysis; it has no network access or file-writing permissions.
- Sanitization: No input sanitization or validation of the processed code content is described.
Audit Metadata