skill-creator

Skill development for AI agents requires structuring components like SKILL.md, scripts, and references. You must define the core task, set degrees of freedom for the agent, and bundle necessary API documentation to ensure effective automated execution.

Progressive disclosure in skill development is a core principle that structures AI agent knowledge to provide concise initial guidance while revealing deeper details only when contextually required, preventing information overload.

When developing a skill, package API documentation by placing the necessary reference files into the references directory, ensuring the AI agent can access specialized knowledge and workflows during execution.

The structure of an AI skill consists of a main SKILL.md file for core instructions, alongside scripts, references, and assets directories that bundle workflows and specialized knowledge for the agent.

Best practices for writing a SKILL.md file include maintaining conciseness, setting appropriate degrees of freedom for the AI agent, and applying progressive disclosure to manage workflow complexity effectively.

Using scripts when building an AI skill is optional but recommended for encapsulating complex logic. Scripts act as executable components that the AI agent can trigger to perform specific workflow automation tasks.