From roach
Use when creating git commits with user approval, whether committing explicit file paths or auto-discovering all changes
How this skill is triggered — by the user, by Claude, or both
Slash command
/roach:committingThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
You are tasked with creating git commits. You support two modes: explicit file paths as arguments, or auto-discovery when no arguments are given.
You are tasked with creating git commits. You support two modes: explicit file paths as arguments, or auto-discovery when no arguments are given.
If file paths are provided as arguments:
Read scoped diffs:
git diff -- <file1> <file2> ... scoped to ONLY the provided filesgit status, git diff, or git diff --cachedAnalyze and split intelligently:
Present your plan to the user:
Execute upon confirmation:
git commit <file1> <file2> ... -m "message"git add followed by git commit -m (this commits ALL staged files!)git log --oneline -n [number of commits created]If no file paths are provided:
Discover all changes:
git status to list all staged and unstaged filesgit diff --cached to inspect staged changesgit diff to inspect unstaged changesAnalyze and propose grouping:
"note: file.ts has both staged and unstaged changes — this will commit only the staged portion"
Present the proposal — always required:
I found changes across N files. Here's how I'd group them:
Commit 1: "feat: add X"
- path/to/file1.ts
- path/to/file2.ts
Commit 2: "chore: update config"
- config/settings.json
Does this grouping look right, or would you like to adjust?
Wait for user confirmation. The user may approve as-is or request adjustments (merge commits, exclude files, reword messages, etc.). Do NOT proceed until confirmed.
Execute upon confirmation:
git commit <file1> <file2> ... -m "message" per commitgit add followed by git commit -mgit log --oneline -n [number of commits created]npx claudepluginhub stefanfaur/roach-marketplace --plugin roachGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.