What problem does it solve?
This Skill helps you convert a messy working tree into a set of logically grouped, well-written commits that a reviewer can understand quickly and safely.
Core Features & Use Cases
- Auto-detect commit message style: Detects the repository's commit convention (Conventional Commits, plain imperative, ticket-prefixed, etc.) by checking policy files and recent history.
- Group changes by intent: Splits unrelated work into separate commits while keeping tightly coupled changes together, including tests and wiring.
- Write reviewer-friendly messages: Produces a commit subject plus a rationale-focused body (Why/Before/After/Findings as appropriate) and surfaces issue links when available.
- Stages and commits safely: Uses explicit file lists for
git add, avoids committing secrets, and asks for confirmation when boundaries are ambiguous.
Quick Start
Ask the AI to commit your current changes with commit messages that match this repository’s style, splitting unrelated work into separate commits and including reviewer-oriented explanations and any relevant issue links.