What problem does it solve? AI agents generate code at high speed, and without disciplined version control, changes become unreviewable, irreversible, and untraceable. This Skill enforces atomic commits, short-lived branches, semantic versioning, and curated changelogs so every change stays manageable and every release communicates clearly to consumers. ## Core Features & Use Cases - Commit Discipline: Enforces atomic commits with conventional message types (feat, fix, refactor, test, docs, chore), pre-commit hygiene checks, and the save-point pattern so no more than one increment of work is ever lost. - Branching & Parallel Work: Provides trunk-based development guidance, branch naming conventions, and git worktree workflows so multiple agents can work in parallel without interfering. - Release & Versioning: Defines semantic versioning rules (MAJOR/MINOR/PATCH), tag-as-source-of-truth releases, and human-readable changelogs grouped by Added/Changed/Fixed/Deprecated/Removed/Security. - Use Case: An agent finishes a feature slice, runs tests and lint, commits with a descriptive conventional message, and when cutting a release, tags v1.4.0 and writes a consumer-facing changelog entry in the same change. ## Quick Start Ask the agent to commit the current changes as atomic commits with conventional messages and verify no secrets are staged.