Skip to content

CONTRIBUTING.md

Latest commit

 

History

History
119 lines (90 loc) · 4.01 KB

CONTRIBUTING.md

File metadata and controls

119 lines (90 loc) · 4.01 KB
Raw

🤝 Contributing to Claude Skills

First off, thank you for considering contributing to Claude Skills! Every skill you add helps thousands of developers work smarter with AI.

📋 Table of Contents

Ways to Contribute

Contribution Description
🆕 New Skill Create a brand new skill file
🔧 Improve Existing Enhance a skill's prompt, add examples, or fix edge cases
📝 Documentation Fix typos, improve explanations, or add translations
🧪 Testing Test skills and report results or issues
💡 Skill Ideas Open an issue with a skill idea for others to implement

Submitting a New Skill

1. Check for Duplicates

Search the existing skills to make sure your idea isn't already covered.

2. Use the Template

Copy the SKILL_TEMPLATE.md file and fill in all sections:

cp SKILL_TEMPLATE.md skills/<category>/your-skill-name.md

3. Choose the Right Category

Category Use When...
developer/ Code-related tasks (review, generation, debugging)
security/ Security analysis, vulnerability scanning, threat modeling
data/ Data analysis, SQL, visualization, dataset manipulation
writing/ Documentation, content creation, communication
architecture/ System design, patterns, scalability
devops/ CI/CD, containers, infrastructure, deployment
productivity/ Workflow optimization, communication, learning
creative/ Design, naming, brainstorming, ideation

4. Test Your Skill

Before submitting, test your skill by:

  • Pasting the system prompt into Claude.ai
  • Running at least 3 different inputs (easy, medium, edge case)
  • Verifying the output matches your documented examples

Quality Standards

Every skill must have:

  • A clear, one-line description
  • Complete metadata (category, difficulty, compatibility, token estimate)
  • A well-structured system prompt using XML tags
  • At least 2 example interactions with realistic inputs/outputs
  • Tips & Variations section with at least 2 modification ideas
  • System prompt under 1500 tokens (check with token counter)

Every skill should have:

  • Edge cases documented
  • Error handling instructions in the prompt
  • Clear scope boundaries (what it does AND doesn't do)

Pull Request Process

  1. Fork the repository
  2. Create a new branch: git checkout -b skill/your-skill-name
  3. Add your skill file in the correct category folder
  4. Update README.md to add your skill to the catalog table
  5. Commit with a descriptive message: feat: add sql-query-optimizer skill
  6. Push and open a Pull Request

PR Description Template

## New Skill: [Skill Name]

**Category**: [e.g., Developer]
**Difficulty**: [e.g., ⭐⭐]

### What does this skill do?
[Brief description]

### Why is it useful?
[Real-world use case]

### Testing done
- [ ] Tested on Claude.ai
- [ ] Tested with Claude Code
- [ ] Tested via API
- [ ] Included example I/O in skill file

Style Guide

File Naming

  • Use kebab-case: code-reviewer.md, threat-modeler.md
  • Be descriptive but concise
  • No version numbers in filenames

Prompt Writing

  • Use XML tags for structure: <role>, <goal>, <constraints>, <output_format>
  • Write in second person: "You are...", "Your task is..."
  • Be specific over vague: "List exactly 3 issues" > "List some issues"
  • Include uncertainty handling: Tell Claude what to do when unsure

Markdown Formatting

  • Use proper heading hierarchy (# → ## → ###)
  • Include a table of contents for skills longer than 100 lines
  • Use code blocks with language identifiers

Thank you for making Claude Skills better for everyone! 🧠✨