What problem does it solve?
This Skill provides a structured, repeatable workflow for building and refining Claude Code skills. It guides users through defining a goal, drafting SKILL.md with a frontmatter name and description, creating eval prompts, running iterative trigger evaluations, and using results to improve skill descriptions and implementation guidance. It also supports bundling resources (scripts, references, assets) to automate common tasks and accelerate iteration.
Core Features & Use Cases
- Capture Intent: define what the skill enables and when it should trigger.
- Interview and Research: ask clarifying questions to align edge cases and success criteria.
- Write the SKILL.md: populate YAML frontmatter (name, description) and craft the body with actionable steps, constraints, and usage guidelines.
- Progressive Disclosure and Bundled Resources: attach optional scripts, references, and assets for on-demand execution and context loading.
- Evaluation Loop: generate eval sets, run trigger evaluations, analyze results, iterate until the skill description and behavior are robust; optionally run optimization of the description.
Quick Start
Draft the skill by defining its goal, create SKILL.md with frontmatter, build a minimal eval set, run the evaluation loop, and apply one improvement pass.
