Claude Code Skill Generator

Step 1: Fetch Latest Documentation

1. Read the Agent Skills overview: https://docs.claude.com/en/docs/agents-and-tools/agent-skills/overview
2. Read the best practices guide: https://docs.claude.com/en/docs/agents-and-tools/agent-skills/best-practices
3. Read the Claude Code skills guide: https://docs.claude.com/en/docs/claude-code/skills
4. Read the quickstart guide: https://docs.claude.com/en/docs/agents-and-tools/agent-skills/quickstart
5. Extract the complete "Checklist for effective Skills" section from the best practices guide

Achieve these tasks as a checklist, printing progress as you go:

• Fetched Agent Skills overview documentation
• Fetched best practices documentation
• Fetched Claude Code skills guide
• Fetched quickstart guide
• Extracted quality checklist from documentation
• Ready to help create skills

Once complete, notify the user that you're ready to help create their skill.

Step 2: Understand the Skill Requirements

1. "What skill are you developing today?"

   • What specific task or objective does this skill address?
   • What domain expertise should it provide?
   • When should Claude trigger this skill?

2. "Do you envision this skill evolving over time?"

   • Will this skill grow in complexity?
   • Do we need a multi-file strategy from the start?
   • Consider:
     • Single file: Simple, focused skills under 500 lines
     • Multi-file: Complex skills with multiple domains, workflows, or extensive reference materials

Wait for their responses before proceeding to Step 3.

Step 3: Create Implementation Plan

Implementation Tasks

• Design skill structure (single-file vs multi-file)
• Write YAML frontmatter (name and description)
• Create SKILL.md with core instructions
• Add workflows and examples (if needed)
• Create additional reference files (if multi-file strategy)
• Add utility scripts (if needed for complex operations)

Quality Verification

Use the complete "Checklist for effective Skills" that you extracted from the documentation in Step 1.

Include all checklist items from the official documentation, organized as they appear in the docs (Core Quality, Code and Scripts, Testing sections).

Testing Plan

Create a testing plan with:

1. Test Scenarios: Minimum 3 realistic use cases that exercise the skill
2. Expected Behavior: What should happen when the skill is triggered
3. Success Criteria: How to verify the skill works correctly
4. Edge Cases: Potential failure modes to test

Present this plan to the user for approval before implementing.

Step 4: Implement the Skill

1. Create the skill structure based on single-file or multi-file strategy
2. Write YAML frontmatter with properly formatted name and description
3. Implement core instructions in SKILL.md body
4. Add supporting files if using multi-file strategy
5. Create utility scripts if needed
6. Add examples and workflows to guide Claude

Follow these principles from the documentation:

• Concise is key: Assume Claude is smart, only add what it doesn't know
• Set appropriate degrees of freedom: Match specificity to task fragility
• Use progressive disclosure: Reference files for detailed content
• Implement workflows: Provide clear, sequential steps for complex tasks
• Add feedback loops: Include validation steps for quality-critical operations

Step 5: Verify Quality

1. Go through Core Quality checklist items
2. Go through Code and Scripts checklist items (if applicable)
3. Report any items that need attention
4. Make necessary corrections

Present verification results to the user.

Step 6: Test the Skill

1. Run each test scenario defined in the plan
2. Verify expected behavior for each test
3. Document results: What worked, what needs adjustment
4. Test edge cases and failure modes
5. Iterate if needed: Refine the skill based on test results

Present test results to the user with recommendations for improvements if needed.

Step 7: Finalize and Deliver

1. Show the complete skill structure (all files)
2. Provide installation instructions for Claude Code
3. Document usage examples showing how to trigger the skill
4. Summarize what the skill does and its key features
5. Suggest potential improvements for future iterations

Key Principles

Throughout the process, adhere to these principles:

• Be concise: Every token matters; assume Claude's intelligence
• Be specific: Clear, actionable instructions over vague guidance
• Be systematic: Follow the checklist rigorously
• Be user-focused: Ask questions when requirements are unclear
• Test thoroughly: Verify the skill works in realistic scenarios