agentleFS
jordantplowsjordantplows/claude-skills/.claude/skills/commit-writer/SKILL.md

commit-writer

Skill1 starsChanged 5 months agoMITView on GitHub

Writes a complete conventional commit message from currently staged changes. Analyzes the diff, determines the correct type (feat/fix/ refactor/chore/docs/test), writes a concise summary line, and a full body explaining what changed and why.

2 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions2 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 2

    1. megamott/vibe-skills★ 28same name
    2. ivanestar/eitr★ 4same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository3 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

12 rules. About 0 are common, 12 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈12

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If nothing is staged, inform the user and exit.

    line 21
  2. Distinctive:

    If the diff is very large (>500 lines), sample key files instead:

    line 27
  3. Distinctive:

    test: test additions or changes

    line 46
  4. Distinctive:

    If changes span multiple types, pick the dominant one and mention others in the body.

    line 50
  5. Distinctive:

    No period at the end

    line 56
  6. Distinctive:

    Breaking changes: Call out explicitly with "BREAKING CHANGE:" on its own line, followed by what breaks and how to migrate.

    line 95
  7. Distinctive:

    User context: If the user provided extra context as an argument to this skill, weave it naturally into the body rather than appending.

    line 98
  8. Distinctive:

    If staged changes span genuinely unrelated concerns (e.g., a feature plus an unrelated dependency bump), flag this:

    line 103
  9. Distinctive:

    Wait for their answer before proceeding.

    line 113
  10. Distinctive:

    Ask: "What should I change?" Wait for feedback, regenerate, show again.

    line 150
  11. Distinctive:

    Wait for feedback, regenerate, show again.

    line 151
  12. Distinctive:

    Always append this line to the commit body:

    line 159

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

commit-writer

Generates proper conventional commit messages from staged changes.

Instructions

  1. Check what's staged
    git diff --staged --stat
    If nothing is staged, inform the user and exit.
  2. Read the actual changes
    git diff --staged
    If the diff is very large (>500 lines), sample key files instead:
    git diff --staged --stat
    git diff --staged path/to/key/file.ts
  3. Read recent commit history for style
    git log --oneline -10
    git log -1 --format="%B"
    This shows the repository's commit message conventions.
  4. Determine the commit type

    • feat: new functionality added
    • fix: bug fix
    • refactor: restructuring without behavior change
    • chore: tooling, config, dependencies, build setup
    • docs: documentation only
    • test: test additions or changes
    • perf: performance improvement
    • style: formatting only, no logic change

    If changes span multiple types, pick the dominant one and mention others in the body.

  5. Write the subject line Format: type: short imperative summary
    • Under 72 characters
    • No period at the end
    • Imperative mood ("add", "fix", "merge" — not "added", "fixed", "merged")
    • Be specific but concise
  6. Decide if body is needed

    Skip the body for:

    • Typo fixes
    • Single small bug fixes
    • Dependency bumps
    • Self-explanatory one-liners

    Write a full body for:

    • Structural changes
    • Multi-file refactors
    • New features
    • Anything that changes behavior
    • Anything where "why" isn't obvious from "what"
  7. When writing a body

    Structure:

    First paragraph: what changed and why, in plain English.
    
    Changes:
    - Group related changes logically
    - Use bullet points for clarity
    - Include file paths or component names
    
    Updated:
    - List modified existing features
    - Explain impact on behavior
    
    Renamed:
    - Old → new mappings
    - Explain why renamed

    Breaking changes: Call out explicitly with "BREAKING CHANGE:" on its own line, followed by what breaks and how to migrate.

    User context: If the user provided extra context as an argument to this skill, weave it naturally into the body rather than appending.

  8. Check for unrelated changes

    If staged changes span genuinely unrelated concerns (e.g., a feature plus an unrelated dependency bump), flag this:

    "These look like two unrelated changes:

    1. [feature description]
    2. [dependency/fix description]

    Want me to write one message covering both, or should we split into separate commits?"

    Wait for their answer before proceeding.

  9. Show the message and confirm

    Display the complete commit message in a code block, then ask:

    Use this commit message?
    [Y] yes  [E] edit  [N] cancel
  10. On [Y] — commit

    For short messages:

    git commit -m "type: subject" -m "body paragraph"

    For complex multi-paragraph bodies, write to a temp file:

    cat > /tmp/commit-msg.txt <<'COMMIT_EOF'
    type: subject
    
    First paragraph explaining what and why.
    
    Changes:
    - item one
    - item two
    
    Co-Authored-By: Claude Sonnet 4.5 (1M context) <noreply@anthropic.com>
    COMMIT_EOF
    git commit -F /tmp/commit-msg.txt
    rm /tmp/commit-msg.txt
  11. On [E] — edit

    Ask: "What should I change?" Wait for feedback, regenerate, show again.

  12. On [N] — cancel

    Confirm cancellation and exit without committing.

Co-Author tag

Always append this line to the commit body:

Co-Authored-By: Claude Sonnet 4.5 (1M context) <noreply@anthropic.com>

Examples

Example 1: Feature addition

Staged changes: New authentication flow with JWT tokens

Message:

feat: add JWT-based authentication system

Replaces session cookies with JWT tokens for better scalability
and easier mobile client support. Tokens expire after 7 days
and include refresh token rotation.

Changes:
- src/auth/jwt.ts — token generation and verification
- src/middleware/auth.ts — JWT validation middleware
- src/routes/auth.ts — login/logout/refresh endpoints
- package.json — added jsonwebtoken dependency

Updated:
- User model to store refreshTokenVersion for invalidation
- API clients to include Authorization header

BREAKING CHANGE: clients must now send 'Authorization: Bearer <token>'
header instead of relying on cookies. See migration guide in docs/auth.md.

Co-Authored-By: Claude Sonnet 4.5 (1M context) <noreply@anthropic.com>

Example 2: Bug fix (no body needed)

Staged changes: One-line fix for null pointer

Message:

fix: prevent null pointer in user profile loader

Example 3: Refactor

Staged changes: Reorganized directory structure

Message:

refactor: move executive agents into department directories

Executives now live inside the department they steer instead of
as separate top-level folders. This reflects that an executive
is the steering layer of its org, not a standalone entity.

Changes:
- cfo/ → finance/exec/
- cmo/ → marketing/exec/
- cto/ → engineering/exec/
- cpo/ → product/exec/
- coo/ → operations/exec/
- All department agents updated with reportsTo field

Updated:
- company-os.ts schema — exec state nests under department
- router.ts — agent discovery updated for new structure
- All .md agents — corrected read/write paths

Renamed for clarity:
- product/exec/roadmap.md → product/exec/roadmap-oversight.md
- operations/exec/operations.md → operations/exec/daily-operations.md

No functional behavior change — this is structural only. All
agent logic, watch/reason/act cycles, and company.os coordination
remain the same.

Co-Authored-By: Claude Sonnet 4.5 (1M context) <noreply@anthropic.com>

Usage

In any conversation with staged changes:

/commit-writer

Or with context:

/commit-writer This fixes the bug reported in issue #42 where users 
couldn't log in after password reset.

The skill will analyze the diff, generate the message, and walk you through confirmation before committing.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.