skill-refinement

Create an AI skill by following a step-by-step process: define frontmatter metadata, select a structure pattern (Workflow-Based, Task-Based, Reference/Guidelines, or Capabilities-Based), organize bundled resources like scripts and references, validate your SKILL.md file, and package for distribution using provided automation tools.

Design skills using proven patterns for workflows and output formats with progressive disclosure. Structure skills with proper anatomy, standardized components, and appropriate degrees of freedom. Use template generation scripts and validation tooling to ensure maintainability and avoid common pitfalls in skill design.

Yes. Use included automation scripts—init_skill.py initializes new skills with proper structure, quick_validate.py validates SKILL.md format, and package_skill.py creates distributable packages. These tools enable deterministic execution and reliable packaging for your skills.

Package skills with scripts and references as core components. Include frontmatter defining metadata, a validated SKILL.md file documenting structure and patterns, organized resource bundles, and automation scripts that support initialization, validation, and packaging workflows.

Workflow-Based skills orchestrate multi-step processes with sequential logic, while Task-Based skills focus on discrete, independent actions. Reference/Guidelines skills provide knowledge and best practices, and Capabilities-Based skills extend Claude with specialized functions. Choose based on your skill's primary use case.

This skill provides comprehensive guidance suitable for creating modular skills, but familiarity with skill anatomy, structure patterns, and documentation standards helps. The included templates and validation tools reduce barriers by automating structural setup and catching common errors early.