What problem does it solve? Confirming that a change to a browser game actually works requires more than reading the code: builds go stale, service workers serve old bundles, and balance claims need real playthroughs. This Skill encodes the exact procedure for building, serving, driving, and bot-playing this repo's games so verification produces trustworthy results. ## Core Features & Use Cases - Build and serve correctly: Distinguishes the React/TypeScript Payload build from single-file games, rebuilds dist/ before testing, and clears stale service workers that hide new code. - Scripted bot playthroughs: Drives full game runs through accessible names in the browser pane, using lazy and competent bots to validate quota curves and difficulty tuning with real numbers. - Completion gates: Enforces a checklist before calling work done, including clean console errors, full runs to the end screen, seed determinism checks, npm test, npm run typecheck, and a clean git status. - Use Case: After changing Payload's quota math, use this Skill to rebuild, clear the service worker, run both bots to completion, and report actual scores proving Easy is winnable and Hard is appropriately tight. ## Quick Start Ask the assistant to verify that your recent change to the game actually works by building, serving, and bot-playing a full run.
