skill-creator

To create a reusable AI skill, you initialize a new skill folder, populate the SKILL.md file with clear YAML frontmatter and concise body instructions, then run provided scripts to generate the agents/openai.yaml UI metadata and validate the folder structure.

Progressive disclosure in skill development is the practice of keeping SKILL.md body instructions concise while placing deterministic tasks in scripts and domain knowledge in references, preventing context bloat and ensuring consistent behavior.

Yes, PyYAML is a required dependency for generating the agents/openai.yaml file, which contains display_name, short_description, and optional interface fields for consistent UI metadata.

You validate a SKILL.md folder structure by running the provided validation scripts after populating the SKILL.md frontmatter, optional scripts, references, and assets, ensuring the skill conforms to best practices and consistent behavior.

SKILL.md YAML frontmatter must include a name and description at minimum, with the body following progressive disclosure patterns to organize procedural workflows and tool integrations effectively.

Yes, you can add optional scripts and references to a skill folder to handle deterministic tasks and provide domain knowledge, keeping the main SKILL.md instructions concise and reducing context bloat.