atomic-commit
Guide splitting large commits into atomic, focused commits
Best use case
atomic-commit is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Guide splitting large commits into atomic, focused commits
Teams using atomic-commit 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/atomic-commit/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How atomic-commit Compares
| Feature / Agent | atomic-commit | 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?
Guide splitting large commits into atomic, focused commits
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
# Atomic Commit Skill Help users create atomic commits - one logical change per commit. ## Available Operations - **analyze** → Determine if changes should be split - **group** → Group related files together - **suggest** → Recommend commit breakdown - **sequence** → Generate commit sequence plan - **interactive** → Interactive splitting guidance ## Usage Examples ```bash # Analyze if splitting needed /atomic-commit analyze # Group files by type and scope /atomic-commit group strategy:type # Suggest commit breakdown /atomic-commit suggest # Create commit sequence /atomic-commit sequence groups:"feat:5,fix:2,docs:1" # Interactive splitting /atomic-commit interactive ``` ## Atomic Commit Principles **One logical change per commit:** - ✅ Single type (all feat, or all fix) - ✅ Single scope (all auth, or all api) - ✅ Reasonable size (≤10 files) - ✅ Logically cohesive - ✅ Can be reverted independently ## Router Logic Parse $ARGUMENTS to determine operation: 1. Extract first word as operation name 2. Parse remaining parameters as key:value pairs 3. Route to appropriate operation file 4. Handle errors gracefully **Available operations:** - `analyze` → Read `commands/atomic-commit/analyze-splitting.md` - `group` → Read `commands/atomic-commit/group-files.md` - `suggest` → Read `commands/atomic-commit/suggest-commits.md` - `sequence` → Read `commands/atomic-commit/create-sequence.md` - `interactive` → Read `commands/atomic-commit/interactive-split.md` **Error Handling:** - Unknown operation → List available operations with examples - No changes detected → Prompt user to make changes first - Already atomic → Confirm no split needed **Base directory**: `commands/atomic-commit/` **Current request**: $ARGUMENTS ### Processing Steps 1. Parse operation from $ARGUMENTS 2. Validate operation exists 3. Extract parameters 4. Read corresponding operation file 5. Execute operation with parameters 6. Return results with actionable guidance ### Example Flows **Quick analysis:** ``` /atomic-commit analyze → Analyzes current changes → Returns split recommendation with reasoning ``` **Interactive workflow:** ``` /atomic-commit interactive → Guides step-by-step through splitting → Shows groupings, suggests commits, creates plan ``` **Custom grouping:** ``` /atomic-commit group strategy:scope → Groups files by module/scope → Returns groupings for review ```
Related Skills
commit-error-handling
Handle git errors and edge cases gracefully
commit-best-practices
Enforce git commit best practices and workflow guidance
commit-analysis
Analyze git changes to understand nature, scope, and commit type for intelligent message generation
specification-writing
Core reference for EARS requirement patterns, Given/When/Then acceptance criteria, RFC 2119 language, requirement ID conventions, and specification quality checklists. Background knowledge for the spec-writer agent. Preloaded by spec-writer via the skills frontmatter field; not user-invocable.
refine
Unified specification refinement plugin. Single entry point for all spec operations: greenfield architecture pipelines (principles → design → stack → spec → plan), feature spec creation in existing systems, iterative convergence loops, quality reviews, drift detection, finalization, ticket decomposition, traceable updates, and lifecycle archival. Detects pipeline state from existing artifact frontmatter and routes intelligently. Use whenever you need to create, refine, validate, or evolve a technical specification.
validation-orchestrator
Intelligent validation orchestrator with auto-detection and progressive validation workflows
security-scan
Comprehensive security scanning for secrets, vulnerabilities, and unsafe practices
schema-validation
Validate JSON schemas, required fields, and format compliance for marketplaces and plugins
quality-analysis
Deep quality analysis with scoring, recommendations, and actionable reports
documentation-validation
Validate documentation completeness, format, and quality for plugins and marketplaces
best-practices
Enforce OpenPlugins and Claude Code best practices for naming, versioning, and standards compliance
message-generation
Generate conventional commit messages following best practices