writing-skills

To write a SKILL.md file, structure compliant YAML frontmatter and body content, adhering to official definitions and best practices for AI agent instructions. This ensures the skill is discoverable and token-efficient.

Progressive disclosure is an architecture that structures agent knowledge bases by splitting content into reference files when a SKILL.md exceeds token limits, optimizing conciseness and performance.

Package reusable prompts by developing well-structured Skills using markdown, defining clear names and descriptions, and organizing supporting documentation into reference files for optimal agent consumption.

Best practices for frontmatter validation involve using checklists and anti-patterns to enforce quality, conciseness, and discoverability, ensuring the SKILL.md adheres to official AI agent definitions.

Split content into reference files when your SKILL.md exceeds token limits, utilizing progressive disclosure architecture to maintain conciseness optimization and structure the agent knowledge base effectively.