What problem does it solve?
Developers and reviewers struggle with large, mixed-change working trees that make history noisy, hard to review, and difficult to bisect. This Skill analyzes staged, unstaged, and untracked changes and breaks them into a sequence of small, logically grouped commits to create a clean, meaningful git history.
Core Features & Use Cases
- Change discovery: Runs git status, git diff, and git diff --cached to enumerate staged, unstaged, and untracked modifications.
- Logical grouping: Groups edits by concern, feature, type, or chore using clear heuristics to form atomic commit units.
- Safe execution: Produces a numbered commit plan, waits for operator approval, stages specific files or hunks, commits with conventional types and concise messages, and verifies the result with git log.
- Use case: When you finish work across multiple files and want to produce a review-friendly series of commits (feat, fix, refactor) without pushing or amending existing history.
Quick Start
Analyze the current repository and split all uncommitted changes into a numbered plan of atomic commits, then create those commits locally using conventional commit messages.