code-review-skill
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [SAFE]: The skill consists primarily of comprehensive markdown-based review guides and checklists. It promotes secure coding practices such as SQL injection and XSS prevention, and does not contain any hidden or malicious instructions.
- [COMMAND_EXECUTION]: The skill includes a local utility script
scripts/pr-analyzer.py. A detailed review of the Python code confirms it is benign, using only standard library modules to parse git diff text and calculate complexity metrics without performing any unsafe operations or network calls. - [EXTERNAL_DOWNLOADS]: The skill utilizes the
WebFetchtool to access documentation. Analysis of the reference guides shows that all embedded URLs point to trusted technology documentation sites (e.g., MDN, GitHub, official language sites). No references to risky or untrusted third-party servers were found. - [INDIRECT_PROMPT_INJECTION]: As the skill is designed to process untrusted code from pull requests, it inherently possesses an ingestion surface for indirect prompt injection. However, the skill implements a strict four-phase review structure and uses defined severity labels, which provides a strong execution boundary that mitigates the risk of the agent deviating from its primary task.
Audit Metadata