What problem does it solve? Writing an effective agent skill is hard to get right on the first try: descriptions fail to trigger, instructions get ignored under pressure, and there is no systematic way to know whether a change actually improved the skill. This Skill provides a full lifecycle workflow for drafting skills, running with-skill versus baseline evaluations, grading outputs against assertions, and iterating based on quantitative benchmarks and user feedback. ## Core Features & Use Cases - Guided skill authoring: Captures intent through interview questions, then writes SKILL.md files following best practices for names, descriptions, progressive disclosure, and writing style. - Evaluation harness: Spawns parallel with-skill and baseline runs, drafts verifiable assertions, captures timing and token data, and aggregates results into benchmark.json with pass rates and variance analysis. - Review viewer: Launches a browser-based viewer (via eval-viewer/generate_review.py) showing qualitative outputs and quantitative benchmarks, collecting user feedback into feedback.json for the next iteration. - Description optimization and bulletproofing: Improves triggering accuracy and pressure-tests discipline-enforcing skills using rationalization tables and red-flag patterns. - Use Case: You want a skill that formats release notes. The Skill helps you draft it, creates three realistic test prompts, runs them with and without the skill, shows you a side-by-side comparison in the browser, and rewrites the skill based on your feedback until the benchmark shows consistent improvement. ## Quick Start Help me create a new skill that converts meeting transcripts into structured action-item lists, then test and refine it.
