meta-skill-creator

To create an AI skill, follow a structured development process that defines modular, self-contained extensions using a SKILL.md file and bundled resources like scripts, references, and assets. This approach ensures specialized knowledge is efficiently integrated.

Progressive disclosure in AI skill development is a context management principle that structures documentation to present only necessary information initially. It keeps skills concise by loading detailed references or scripts only when the workflow demands them.

Structure scripts and references by organizing them into dedicated directories within your skill bundle. This modular design ensures the AI can access specialized knowledge and executable workflows self-contained, promoting efficient context management.

The best way to document company-specific knowledge is creating a structured skill with a concise SKILL.md file and bundled assets. This modular approach encapsulates specialized workflows and reference materials as self-contained agent extensions.

You need to bundle assets with your AI skill when the workflow requires external files or specialized data to execute tasks. Bundling these resources ensures the skill remains self-contained and can manage context efficiently without missing critical dependencies.