Menu
Run the Stims quality gate. Knows the fast→full iteration loop and visual verification.
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Baseado na classificação ocupacional SOC
Auto-guardrail for agent-authored changes. Use at the start of any coding session to prevent the most common agent mistakes in this repo.
Iterate on visualizer UI, workspace chrome, and CSS with fast feedback loops. Use when changing frontend components, styling, layout, or animation without needing the full MilkDrop runtime.
Modify the flagship visualizer runtime safely. Use when changing the shared runtime, loader, renderer, shell, controls, audio flows, capability checks, or canonical-route behavior.
Review changes to module loading, bootstrap, toy manifest, library resolution, or gamepad polling. Use when a PR touches src/js/bootstrap/, src/js/loader.ts, src/js/router.ts, src/js/toy-view.ts, src/data/toys.json, or gamepad polling code.
Review changes to renderer capability probing, fallback chains, timeout logic, or audio worklet initialization. Use when a PR touches src/js/core/renderer-*, src/js/core/audio-handler.ts, or src/js/milkdrop/runtime/backend-fallback.ts.
Review changes to WebGPU/WebGL dual-backend parity. Use when a PR touches feedback managers, renderer adapters, shader lowering, or any code that affects both WebGPU and WebGL rendering paths.
| name | qa |
| description | Run the Stims quality gate. Knows the fast→full iteration loop and visual verification. |
# 1. Syntax + types + lint (15s)
bun run check:quick
# 2. Targeted test (30-90s)
bun run test tests/path/to/spec.test.ts
# 3. Visual verification
bun run dev
# Open http://localhost:5173/?agent=true
# 4. Full quality gate (before commit, 2-5 min)
bun run check
bun run check:architecture # Cross-boundary import violations
bun run check:toys # Preset manifest drift
bun run check:seo # SEO surface health
bash scripts/check_production_edge.sh # WAF/DNS health