What problem does it solve? Teams working with AI coding agents often end up with inconsistent branch names, mixed-layer commits, and unreviewed changes pushed to the wrong branches. This Skill enforces a strict Git Flow branching model and commit conventions so every AI-assisted change lands in the correct branch with a clean, reviewable history. ## Core Features & Use Cases - Branch Type Diagnosis: Automatically determines whether a task requires a feature, bugfix, hotfix, or release branch, and selects the correct parent branch (develop, main, release, or an existing feature branch). - Conventional Commits with Module Scopes: Generates atomic, layer-specific commit messages in the format <type>(<module>): [ID] <subject>, following the Domain -> Data -> UseCase -> UI flow. - Human Review Gate: Prohibits autonomous commit, push, and merge operations, requiring the user to review and commit changes manually. - Use Case: When an AI agent implements a new feature like a high-temperature push notification, it creates feature/MM-01-high-temp-notification from develop, stages files as it works, and prepares atomic commits for user review. ## Quick Start Ask the agent to start a new feature task and it will diagnose the repository state, propose the correct branch, and prepare conventional commits for your review.
