The Agent Skills Directory

[SAFE]: The skill implements a principle-driven refactoring workflow. All scripts and references point to internal components within the skill's own environment (e.g., ~/.claude/skills/refactor-chain/).
[SAFE]: The checklist.mjs script performs a dynamic import of the principles registry from a fixed local path. This is a standard modular design pattern for loading internal skill configurations and does not involve untrusted input or remote sources.
[SAFE]: While the skill ingests untrusted code as data for analysis, it explicitly mitigates risks by requiring successful build and test passes after every atomic transformation, ensuring no unintended functional changes occur.

refactor-code-principles