- Home
- Skills
- Game Development
- Game QA Scorecard — The Ten Tests Before You Launch
Game QA Scorecard — The Ten Tests Before You Launch
For indie game developers: ten independent tests — errors, looks, feel, fun, balance, audio, speed, screens, ad rules, and the cold first session — and your game doesn't ship until all ten say PASS.
Free
Works with the AI Agents you already use
About this skill
Most games ship on one test: "it runs on my machine." This free scorecard gives your agent ten pre-launch tests to run before your game is allowed to publish:
- Errors (PROOF) — zero errors, and the package is what it claims to be
- Looks (LOOK) — the graphics and covers a player actually sees
- Feel (FLOW) — movement, timing, and how the game responds
- Fun (PLAYER) — the ten-second hook and the one-more-level pull
- Balance (BALANCE) — the difficulty ramp, with no impossible or trivial outliers
- Audio (EAR) — audio files, loops, mute, and ad-pause wiring
- Speed (SPEED) — package weight, load time, and per-frame cost
- Screens (FIT) — every screen shape, mouse and touch
- Ad rules (TILL) — a declined reward can never trigger an ad
- First session (NEWCOMER) — a cold player, taught by doing inside thirty seconds
Each test files its own PASS or FAIL with evidence; any FAIL returns a numbered fix list instead of a vibe. Run it on your build before you submit anywhere — it is the exact scorecard a working game studio applies to its own games.
The full protocol — complete test specifications and a large-scale repeated-session gauntlet method — is the paid companion on this creator's page.
- Ten separate disciplines instead of one "looks fine to me" — including the two most skipped tests: is it actually fun, and can a cold player learn it with no instructions?
- A clear PASS / FAIL bar for every test, so verdicts are decisions, not opinions
- Every verdict filed with evidence and, on a FAIL, a numbered fix list you can work through in order
- No self-grading — when you run the tests in separate agents or sessions, the verdicts stay blind to each other, and a fix re-earns its pass by re-running the test in full
- Catches the expensive failures before submission — broken ad rules, clipped layouts, unreachable targets, and dead first sessions
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Game Development.
Lite version excludes the high-volume stress-testing gauntlet. Verdicts rely on play simulation data provided to the agent. No automated file-level code scanning included.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 12 days ago
- Free with an Agensi account
Permissions
No special permissions declared or detected