skill-builder

Structure a Claude Skill by defining YAML frontmatter with metadata fields (name, description, type, depth, model), then organize content into folders: SKILL.md as the entry point, references/ for deep knowledge, and templates/ for output scaffolds. This ensures consistency, discoverability, and scalability across simple, standard, and complex skill types.

Automate skill creation using Python or TypeScript scripts integrated into your skill structure to handle deterministic tasks, enforce schema validation, and orchestrate external tools. This reduces manual errors and ensures reproducible, standardized skill artifacts across your project.

Yes. Integrate optional Python or TypeScript scripts into your skill's structure to validate frontmatter fields, enforce depth-scaled content limits, and verify folder layouts. This enables deterministic validation workflows that catch inconsistencies before deployment.

Use standard structure for domain-specific capabilities with moderate complexity, organizing content into SKILL.md, references/, and templates/. Choose complex structure when your skill requires extensive external tool orchestration, multiple validation workflows, or advanced automation—both follow the same frontmatter guidelines but scale content depth accordingly.

Define clear YAML frontmatter with keywords, apply progressive disclosure by separating entry-point content from references, use consistent folder organization, and integrate validation scripts. This creates maintainable, discoverable skills that meet project standards and reduce technical debt over time.

Required frontmatter fields include name, description, type (simple, standard, or complex), depth (base, intermediate, or advanced), and model. These fields ensure your skill is properly categorized, indexed, and scaled for content limits while maintaining consistency across your skill portfolio.