skill-creator

To structure an AI skill, you use YAML frontmatter for metadata and markdown instructions for logic, organizing bundled resources like scripts, references, and assets into directories to ensure clean execution.

Best practices for AI skill development include maintaining conciseness, applying progressive disclosure, and setting appropriate degrees of freedom in your prompt engineering to balance strict guidance with flexible AI reasoning capabilities.

You bundle scripts and references by placing them into dedicated directories alongside your main skill file, allowing the AI to access helper scripts and necessary documentation like API guides efficiently during execution.

You need to use progressive disclosure in skill creation when providing large amounts of documentation or complex instructions, ensuring the AI loads detailed context only when required to minimize token usage and maintain conciseness.

Skill creation does not require specific external dependencies or libraries, as the guide focuses on structuring native markdown instructions, YAML frontmatter, and bundled resource assets to extend AI capabilities natively.