refactor-instructions
Pass
Audited by Gen Agent Trust Hub on Apr 22, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill uses basic shell commands such as 'ls' and 'ln' to manage local instruction files like CLAUDE.md and AGENTS.md. These operations are scoped to file discovery and organization within the repository root.
- [SAFE]: The refactoring process includes mandatory user confirmation steps via AskUserQuestion before resolving content contradictions or applying structural changes to the disk, ensuring user oversight and preventing automated file corruption.
- [SAFE]: No remote downloads, external dependencies, or obfuscation techniques were identified in the skill instructions.
- [SAFE]: While the skill processes untrusted instruction files which could potentially contain indirect prompt injections, the human-review requirement for all proposed changes effectively mitigates the risk of the agent executing unintended instructions found within the data.
Audit Metadata