skill-creator

To create an AI agent skill, initialize a skill directory and edit the `SKILL.md` file alongside bundled resources like `scripts/`, `references/`, and `assets/` to ensure modular design and clear documentation.

Best practices for AI agent skill documentation include maintaining conciseness, applying phased disclosure, and following naming conventions to guide the agent effectively without overwhelming it with unnecessary context.

Package AI agent skills by organizing modular components into a distributable format after iteratively refining the `SKILL.md` file and bundled resources based on actual usage and feedback.

You should update an AI agent skill when usage patterns indicate a need for iterative refinement, requiring adjustments to the `SKILL.md` content or bundled resource components to improve efficiency.

Yes, you can use this structured methodology to build skills for specific contexts like cloud deployments by organizing platform-specific deployment scripts and reference documentation within the skill directory.

Phased disclosure in AI agent skill design is a methodology principle that controls the level of freedom and information an agent receives, ensuring it processes tasks efficiently without being overloaded.