skill-creator

To create an AI skill, you structure modular packages into components like SKILL.md, scripts, and references, following best practices for content organization and progressive disclosure to extend AI capabilities.

Progressive disclosure in AI skill development is managing context efficiently by layering information across metadata, the SKILL.md body, and bundled resources to prevent overwhelming the AI.

The best way to structure a self-contained AI skill is using a modular anatomy with a SKILL.md file for core logic, alongside scripts, references, and assets directories for specialized knowledge.

The skill development lifecycle works by guiding you step-by-step from understanding the skill's requirements to structuring its components, setting degrees of freedom, and iterating on the workflow.

You need specific scripts and assets to build effective AI skills when your workflow requires tool integrations or specialized knowledge, which are bundled as self-contained modular units.

You should not use modular skill units for workflow automation if the task cannot be concisely structured into self-contained components or fails to benefit from setting defined degrees of freedom.