Agent-facing clean-code guardrail for handwritten production code, tests, migrations, executable scripts, and code examples intended for use. Use whenever an agent generates or modifies such code, alongside the authorized implementation, bug-fix, refactor, or other code-writing workflow even when the user does not name this skill; explicit `$clean-code` invocation remains supported. Do not use for review-only requests, code explanation without edits, generated or vendored files, or non-code artifacts.
