ThatXliner

Use when committing code — writing commit messages, staging, or reviewing commits before push.

Remove signs of AI-generated writing from text — rhythm, sentence mechanics, contrast frames, and stock AI vocabulary. Use when editing or reviewing text to make it sound more natural and human-written.

Refactor code safely using clean code principles and design patterns. Use this skill whenever the user asks to refactor, clean up, simplify, restructure, modernize, or "improve" existing code; whenever they complain code is messy, hard to read, hard to test, duplicated, or "spaghetti"; whenever they ask which design pattern to apply; and during code review when suggesting structural improvements. Also use after writing new code

Run, test, and eval the xtras plugin skills (x-humanizer, clean-code-refactoring). Use when asked to run a skill, test or verify a skill change, add an eval case, or red-green build a skill. The driver launches each skill headlessly via claude -p in a throwaway workspace and grades the output with deterministic checks.

Use when staging partial file changes, splitting commits, or excluding hunks - non-interactive alternative to `git add -p` for AI agents that cannot use interactive prompts

Use when cutting a new release of gah - bumps the version, updates the changelog, tags, pushes, and creates the GitHub release. Triggers on "new release", "cut a release", "bump version", "release vX.Y.Z".

Generate travel itineraries in the Open Itinerary format. Use when the user asks to plan a trip, create an itinerary, or generate a .oitinerary.kdl file.

Generate travel itineraries in the Open Itinerary format. Use when the user asks to plan a trip, create an itinerary, or generate a .oitinerary.kdl file.

AST-native code transformation — use patchwork CLI instead of sed/regex for structural find/replace/delete/insert operations