What problem does it solve? Writing an effective skill is hard to get right on the first try — descriptions may fail to trigger, instructions may be ambiguous, and there is no built-in way to measure whether a skill actually improves results. This Skill provides a structured create-test-review-iterate loop with quantitative benchmarking so skills are validated against real test prompts rather than guesswork. ## Core Features & Use Cases - Guided Skill Authoring: Interviews the user about intent, trigger contexts, and output formats, then drafts a SKILL.md following progressive disclosure and writing-style best practices. - Parallel Evaluation Runs: Spawns with-skill and baseline subagent runs for each test prompt, captures timing and token data, and grades outputs against assertions. - Benchmark Aggregation & Review UI: Aggregates pass rates, time, and token statistics into benchmark.json/benchmark.md and launches an HTML viewer for qualitative feedback. - Description Optimization: Generates trigger/no-trigger eval queries and runs an automated loop that rewrites the skill description to maximize triggering accuracy on held-out test queries. - Use Case: A user wants a skill that formats weekly sales reports. This Skill drafts it, runs three realistic test prompts with and without the skill, shows a side-by-side benchmark, collects feedback, and iterates until the outputs pass all assertions. ## Quick Start Ask the assistant to create a new skill for your workflow and let it draft, test, and benchmark the skill iteratively.
