learn
Pass
Audited by Gen Agent Trust Hub on Apr 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a robust 'propose-and-verify' workflow where all generated content must be explicitly reviewed and approved by the user via the AskUserQuestion tool before any files are written to the disk.
- [COMMAND_EXECUTION]: Employs standard shell utilities (find, ls, cat, wc, mkdir) for project reconnaissance and directory management. These commands are executed within the local project scope and do not involve shell piping from remote sources or privilege escalation.
- [DATA_EXPOSURE]: Accesses project-specific configuration and metadata files (such as package.json, CLAUDE.md, and .cursorrules) to identify existing conventions. It does not target sensitive directories like .ssh or access environment variables containing credentials.
- [PROMPT_INJECTION]: Although the skill processes untrusted codebase data which could contain indirect injection attempts, the risk is mitigated by the orchestrator's validation logic, which filters for impact and requires evidence from multiple files, followed by mandatory human review of the final output.
Audit Metadata