What problem does it solve? Writing effective agent skills is hard to do blind — descriptions may not trigger, instructions may underperform, and there is no feedback loop. This Skill provides a structured draft-test-review-improve workflow with quantitative benchmarks so skills are validated against real prompts before shipping. ## Core Features & Use Cases - Guided skill authoring: Interview-driven capture of intent, trigger conditions, and output formats, then generation of a well-structured SKILL.md following progressive disclosure and writing-style best practices. - Evaluation harness: Runs test prompts with and without the skill via parallel subagents, grades outputs against assertions, and aggregates pass rates, timing, and token usage into benchmark.json and benchmark.md. - Description optimization: Generates realistic should-trigger and should-not-trigger queries, reviews them in an HTML template, and runs an automated loop (run_eval.py + improve_description.py) that rewrites the description and selects the best by held-out test score. - Packaging: Validates frontmatter and bundles the skill folder into a distributable .skill zip file. - Use Case: A user says "turn this workflow into a skill." The Skill drafts the SKILL.md, creates test prompts, runs with-skill vs baseline comparisons, presents results in a browser viewer, iterates on feedback, and finally optimizes the trigger description. ## Quick Start Ask the agent to create a new skill for your workflow and run its test cases with benchmark comparison.
