sample-skill

You create reusable AI skills for local agents by drafting SKILL.md files with required frontmatter, validating the schema, and packaging the directory into a distributable artifact through an iterative authoring workflow.

A structured AI skill authoring workflow provides a repeatable template and frontmatter governance, enforcing required fields like name and description while tracking optional components such as scripts, references, and assets.

You validate and package AI skills by enforcing the frontmatter schema to ensure required fields are present, tracking optional components, and preparing the validated directory into a distributable .skill artifact.

Yes, you can benchmark and tune AI skill descriptions iteratively across skill versions, applying description tuning and benchmarking workflows to improve agent performance within local environments.

Required frontmatter fields for a valid AI skill package include name and description, enforced by a validation workflow that also tracks optional components like scripts, references, and assets.

Yes, this skill authoring workflow operates with no external dependencies, applying directly to shell-based hosts and agent environments to draft, validate, and package reusable AI skills locally.