full-repo-review
Pass
Audited by Gen Agent Trust Hub on Aug 31, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses
findto discover source files andpython3to run a local scoring script (~/.claude/scripts/score-component.py). These are standard operations for a codebase analysis tool and are scoped to the project environment. - [SAFE]: The skill implements a multi-wave review process (Wave 0-3) that delegates to a
comprehensive-reviewpipeline. This is a structured approach to code analysis that prioritizes evidence-based findings over automated actions. - [SAFE]: All identified findings are aggregated into a local Markdown report (
full-repo-review-report.md) rather than being sent to an external endpoint, preventing data exfiltration risks. - [SAFE]: The instructions include a 'GATE' mechanism where structural scores from deterministic scripts must be manually confirmed by a reviewer with file-and-line evidence before being added to the final report, reducing false positives and ensuring human-in-the-loop validation.
Audit Metadata