What problem does it solve? Writing an OpenCode skill that triggers reliably and produces good output is hard to get right on the first try. This Skill guides you through a structured create-test-evaluate-iterate loop so skills are validated against real prompts and quantitative benchmarks instead of guesswork. ## Core Features & Use Cases - Guided Skill Authoring: Runs a mandatory intake interview, then drafts SKILL.md files with proper frontmatter, progressive disclosure structure, and staged temp-directory workspaces. - Paired Eval Runs & Benchmarking: Spawns with-skill and baseline runs in parallel, grades outputs against assertions, aggregates pass rates, timing, and token usage into benchmark reports, and serves an interactive review viewer. - Description Optimization: Generates realistic trigger/no-trigger eval queries and runs an automated optimization loop with train/test splits to improve skill triggering accuracy. - Use Case: You want a skill that formats weekly sales spreadsheets. This Skill interviews you about the workflow, drafts the skill, runs test prompts with and without it, shows you side-by-side results in a browser viewer, and iterates until the outputs meet your bar. ## Quick Start Ask the agent to help you create a new OpenCode skill for your workflow and let it interview you, draft the skill, and run evaluation loops on your test prompts.
