What problem does it solve? Writing a reusable AI skill that triggers reliably and produces consistent results is hard: descriptions under-trigger, instructions overfit to test cases, and drafts go untested. This Skill guides you through the full lifecycle of drafting, testing, evaluating, and refining SKILL.md files. ## Core Features & Use Cases - Guided Skill Authoring: Interviews you about intent, trigger contexts, and output formats, then drafts a complete SKILL.md with proper frontmatter (name, description, license) and imperative-style body instructions. - Test-Driven Iteration: Generates realistic test prompts, runs the draft skill against them, reads transcripts for wasted effort, and refines instructions without overfitting. - Description Optimization: Builds trigger-eval query sets (should-trigger and near-miss should-not-trigger) and iterates the description field against held-out test queries for accurate activation. - Use Case: You just walked an AI through a multi-step workflow and want to reuse it. Ask to turn the conversation into a skill, and it extracts the steps, drafts the SKILL.md, tests it on realistic prompts, and iterates until it works reliably. ## Quick Start Turn the workflow we just completed into a reusable skill and test it on a few realistic prompts.
