What problem does it solve? Frontend web apps often contain latent bugs that survive casual code review: JavaScript errors, broken drag-and-drop, localStorage corruption crashes, XSS vectors, collision-detection flaws in canvas games, and cosmetic stubs in desktop simulators. This Skill provides a systematic, layered audit methodology that combines reading the source with exercising every interaction path in a real browser. ## Core Features & Use Cases - Six-layer audit methodology: static source analysis, initialization checks, user interaction paths, data persistence, drag-and-drop quirks, and edge cases/security (XSS, accessibility, race conditions). - Specialized audit tracks: dedicated checklists for HTML5 Canvas games (state machines, collision splice bugs, scaling curves, Web Audio API) and desktop environment simulators (window managers, Minesweeper/Solitaire logic, cosmetic stub detection). - Runtime state probing: techniques for inspecting canvas game internals via browser console expressions, plus worked case studies (Space Invaders, Pac-Man, Flappy Bird, Kanban board, Win95 simulator) showing real bugs found. - Use Case: Before shipping a single-file HTML game or dashboard, run this audit to catch critical issues like collision loops that skip bullets, wave advancement after player death, or drop handlers that crash on external drags. ## Quick Start Ask the AI to audit your HTML web app for bugs using the web-app-audit skill, serving it locally with python3 -m http.server first.