de-slop

To clean AI artifacts from code diffs before a pull request, you can run a dry-run to list issues like scratch notes and filler comments, then apply the selected edits to remove the noise while maintaining code behavior.

AI noise in a code diff includes pointless scratch markdown, redundant comments, and defensive checks abnormal for trusted paths. You remove it by running a cleanup process that targets these artifacts and flattens deeply nested code with early returns.

Yes, code cleanup removes AI-generated scratch notes like NOTES, PLAN, and TODO files while explicitly protecting essential documentation like README.md, CONTRIBUTING.md, CHANGELOG.md, and docs/** unless you provide explicit confirmation to delete them.

The best way to prepare PRs generated by AI tools is to use a diff cleanup flow that removes mock-heavy tests, fake metrics, and type-bypassing casts, favoring small, focused edits that enforce consistency with the surrounding codebase.

Diff cleanup maintains code behavior when removing AI artifacts unless actual fixes are required. It focuses on deleting pointless scratch markdown and inconsistent comments, favoring small, focused edits rather than altering functional logic.