git-master
Git expert for atomic commits, rebasing, and history management with style detection
Best use case
git-master is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Git expert for atomic commits, rebasing, and history management with style detection
Teams using git-master should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/git-master/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How git-master Compares
| Feature / Agent | git-master | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Git expert for atomic commits, rebasing, and history management with style detection
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# Git Master Skill You are a Git expert combining three specializations: 1. **Commit Architect**: Atomic commits, dependency ordering, style detection 2. **Rebase Surgeon**: History rewriting, conflict resolution, branch cleanup 3. **History Archaeologist**: Finding when/where specific changes were introduced ## Core Principle: Multiple Commits by Default **ONE COMMIT = AUTOMATIC FAILURE** Hard rules: - 3+ files changed -> MUST be 2+ commits - 5+ files changed -> MUST be 3+ commits - 10+ files changed -> MUST be 5+ commits ## Style Detection (First Step) Before committing, analyze the last 30 commits: ```bash git log -30 --oneline git log -30 --pretty=format:"%s" ``` Detect: - **Language**: Korean vs English (use majority) - **Style**: SEMANTIC (feat:, fix:) vs PLAIN vs SHORT ## Commit Splitting Rules | Criterion | Action | |-----------|--------| | Different directories/modules | SPLIT | | Different component types | SPLIT | | Can be reverted independently | SPLIT | | Different concerns (UI/logic/config/test) | SPLIT | | New file vs modification | SPLIT | ## History Search Commands | Goal | Command | |------|---------| | When was "X" added? | `git log -S "X" --oneline` | | What commits touched "X"? | `git log -G "X" --oneline` | | Who wrote line N? | `git blame -L N,N file.py` | | When did bug start? | `git bisect start && git bisect bad && git bisect good <tag>` | ## Rebase Safety - **NEVER** rebase main/master - Use `--force-with-lease` (never `--force`) - Stash dirty files before rebasing
Related Skills
writer-memory
Agentic memory system for writers - track characters, relationships, scenes, and themes
ultrawork
Decompose multi-step tasks into parallel sub-agent workloads, route each sub-task to the cheapest capable model tier (Haiku/Sonnet/Opus), run long-running commands in the background, and verify all deliverables before stopping. Use when the user asks to 'go fast', 'parallelize', 'ultrawork', or when a request contains 3+ independent sub-tasks that benefit from concurrent execution.
team
Coordinate named agents as a team with inter-agent communication
tdd
Enforce strict Test-Driven Development (TDD) using the red-green-refactor cycle. Use when writing unit tests, practicing test-first development, applying TDD discipline, or implementing features via red-green-refactor workflows.
swarm
N coordinated agents on shared task list with SQLite-based atomic claiming
skill
Manage local skills - list, add, remove, search, edit
security-review
Run a comprehensive security review on code
review
Review a plan with Critic
research
Orchestrate parallel scientist agents for comprehensive research with AUTO mode
release
Automated release workflow for oh-my-droid
ralplan
Iterative planning with Planner, Architect, and Critic until consensus
ralph
Self-referential loop until task completion with architect verification