review
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill's primary function is local code analysis. No malicious patterns, such as data exfiltration, obfuscation, or persistence mechanisms, were detected.
- [COMMAND_EXECUTION]: The skill utilizes common version control commands (
git diff,git log,git rev-parse) to identify changes and project history. These commands are used strictly for data retrieval relevant to the code review process. - [PROMPT_INJECTION]: The skill demonstrates a design that mitigates indirect prompt injection risks when processing untrusted code diffs:
- Ingestion points: Git diff output and commit logs are retrieved in step 1 and passed to sub-agents in step 3.
- Boundary markers: The instructions for sub-agents (Standards, Spec, and Architecture briefs) use specific delimiters and quoting (e.g., "Report — per file/hunk —", "Quote the hunk") to distinguish between instructions and data.
- Capability inventory: The skill's capabilities are restricted to text analysis and reporting using the
general-purposesub-agent; it does not possess the ability to write to the filesystem, execute arbitrary code from the diff, or communicate over the network. - Sanitization: The process relies on the LLM's internal handling of quoted hunks and structured prompts to prevent execution of instructions embedded in the code being reviewed.
Audit Metadata