Menu
Launch a browser-based UAT tester for the NeoMud web marketplace and WASM game client via Playwright
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Basé sur la classification professionnelle SOC
Launch an AI playtester to play NeoMud via the WebSocket relay and report findings
Launch a QA tester to systematically find bugs, test edge cases, and audit game systems. Use when you want thorough bug-hunting beyond what playtesting covers.
Evaluate architecture decisions, propose infrastructure, assess scalability, and make technology recommendations across the full stack — from hardware to cloud to application. Use when planning new services, scaling existing ones, evaluating hosting/cloud options, designing data pipelines, or making build-vs-buy decisions.
Deploy latest code to staging, wait for CI/CD pipelines, and verify health. Use after committing changes that need to go live on staging.
Audit user data channels, PII handling, abuse vectors, and content moderation across NeoMud Platform and game servers. Zero-trust posture — all inputs are invalid until verified.
Evaluate UX flows, interactions, and conventions across NeoMud marketplace, game client, and maker. Brings real-world UX expertise from industry-leading games and platforms.
| name | web-uat-test |
| description | Launch a browser-based UAT tester for the NeoMud web marketplace and WASM game client via Playwright |
| context | fork |
| agent | web-uat-tester |
Launch a BrowserKnight session using the Playwright MCP browser tools. The agent will open a real browser, navigate the marketplace and WASM game client, take screenshots, play the game, and file bugs from a player's perspective.
$ARGUMENTS
If arguments were provided above, focus your testing session on that specific area. Examples:
marketplace — test only the React marketplace (landing, search, detail, responsive)play transition — test the flow from marketplace → WASM game loading → gamegame combat — get into the game and test combat flowfull journey — complete end-to-end: browse → select world → play → fight → rateIf no arguments were provided, run a full journey exploratory session.
Check if the marketplace dev server is running:
curl -s http://localhost:3000 > /dev/null 2>&1 && echo "Marketplace is running on :3000" || echo "Marketplace is NOT running"
If not running, inform the user:
The marketplace dev server is not running. Please start it:
cd NeoMud-Platform/web && npm run dev
Check if the game server is running:
curl -s http://localhost:8080/health 2>&1
If the game server is not running and the user wants to test the play flow, inform them:
The game server is not running. Please start it:
./gradlew packageWorld --rerun-tasks
./gradlew :server:run
Check if the Platform API is running (needed for marketplace → play flow):
curl -s http://localhost:3002/api/v1/health 2>&1
If the Platform API is not running, the marketplace will show errors loading worlds. For marketplace-only visual testing this is acceptable (test error states). For play-through testing, inform the user:
The Platform API is not running. Please start it:
cd NeoMud-Platform && npm run dev
Ensure the web-uat label exists:
gh label create web-uat --color 1D76DB --description "Web UAT testing findings" 2>/dev/null || true
http://localhost:3000 using browser_navigateAfter the session is complete (after producing the report), you MUST run these cleanup commands:
# Delete all screenshots cached during the session
rm -rf .playwright-mcp/*
rm -f web-uat-*.png