Borrowing it
Nothing to install: this file belongs to gug007/lpm. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/gug007/lpm/main/.claude/skills/short-game/SKILL.mdgit clone --depth 1 https://github.com/gug007/lpmWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/gug007/lpm/short-game)<a href="https://agentmods.dev/skills/gug007/lpm/short-game"><img src="https://agentmods.dev/badge/skills/gug007/lpm/short-game/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/gug007/lpm/short-game"><img src="https://agentmods.dev/badge/skills/gug007/lpm/short-game.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00229 | $0.02776 |
| Opus 5.5 | $0.00092 | $0.01110 |
| Sonnet 5.5 | $0.00046 | $0.00555 |
| Haiku 4.5 | $0.00023 | $0.00278 |
Grade A, and why
short-game scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 91 lines — stays where its author put it; the contents beside it link to each section on GitHub.
A short-race with a game for its prompt. scripts/new.js writes the lesson through short-race's own new.js, so everything in ../short-race/SKILL.md and ../short/SKILL.md still holds (models and CLIs, Run in duplicates, the take, the review) and is not repeated here. This file covers only what a game race adds.
Why games
Creator Search Insights (global, last 7 days, 2026-10-07): "Using AI to build a 3D game" 314K, "making a game with ai" 230K, "ai game generator" 274K, "how to make game with ai" 85K and "ai games to play" 93K (both up 1000%+). For the brand pairs: "claude vs gemini" 267K, "claude ai vs chatgpt" 252K, "claude vs grok" 204K. Races get far more views than feature shorts, and a game that plays itself keeps the reveal moving. The first game race was Opus 5.5 high (12:16) vs Gemini 3.1 Pro (2:24) on the 3D runner.
The games
| key | builds | headline |
|---|---|---|
runner (default) |
a 3D endless runner: lanes, jumps, coins | build a 3D game 🎮 |
racer |
a 3D highway racer weaving through traffic | build a 3D racing game 🏎️ |
flappy |
Flappy Bird through scrolling pipes | build Flappy Bird 🐦 |
snake |
Snake on a 20x20 board | build Snake 🐍 |
shooter |
a space shooter blasting enemy waves | build a space shooter 🚀 |
breakout |
Breakout with fresh brick walls | build Breakout 🧱 |
gta |
a top-down GTA-style police chase in a Vice City-style town | build GTA 6 🚓 |
node scripts/new.js --list prints them. Every prompt in scripts/games.js keeps short-race's rules: index.html at the project root, one file, no libraries or images, a stage with sizes given as shares of it, scaled to fit and never cropped, and "Don't run or test it". On top of that the game plays itself and restarts at once after a crash, so the reveal is live play with nobody at the controls. The stage is 900x700, the shape of a row in the tall window (see One above the other). Races posted before 2026-10-09 ran in columns on a 420x740 stage, and those before 2026-10-10 in short-race's 800x990 window on a wide 900x360 stage: rewrite such a lesson with --force before retaking it. A layout gives each size as a share of the stage (Snake's board spans the height, the shooter's enemies come in from the right). Add a new game to games.js in the same shape (build, play, layout, restart, then the copy fields), not as a one-off --prompt. Keep its whole prompt under about 850 characters: a Claude or Codex copy gets it on a launch line that macOS cuts off at 1024 bytes (see short-race's Traps), and new.js refuses a longer one.
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday Changed · +2 lines a81c52e538ca
- 4d ago Changed · +2 lines bff4a929ea19
- 10d ago Changed 75c9ae6e0d85
- 13d ago Changed · +5 lines · +6 tokens per session cac992091919
- 14d ago Changed · +4 lines · +1 tokens per session 73d1f24f14f4
- 16d ago First seen · 78 lines · 222 tokens per session scan A 8d7e0982e728
short-game is a skill published in the GitHub repository gug007/lpm (145 stars, last pushed today), licensed MIT. It adds 229 tokens to every session and 2,776 once invoked, about $0.0009 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-08.
Other skills, from other repositories
mobile-animation-skia
React Native Skia GPU-accelerated 2D graphics - Canvas, declarative drawing, shaders, image filters, Paragraph text, Atlas batch rendering, Reanimated animations.
web-3d-react-three-fiber
React Three Fiber (R3F) 3D rendering — Canvas, meshes, materials, lights, cameras, animations, events, physics, post-processing, performance.
unreal-engine
Build, debug, test, package, and optimize Unreal Engine projects using C++, Blueprints, Automation Tests, Unreal Build Tool, and Automation Tool. Use for .uproject work and Unreal-specific runtime or editor issues.
game-development
Route and execute game project work across Unreal Engine, Unity, Godot, Bevy, browser engines, assets, performance, builds, and playtesting. Use when a game request spans systems or the engine is not yet confirmed.
game-networking
Design, implement, test, secure, and profile multiplayer game networking across engines, including authority, replication, prediction, rollback, lobbies, reconnects, and adverse network conditions.
bevy
Build, test, debug, and optimize Rust games and tools using the Bevy engine, ECS scheduling, assets, rendering, and platform features. Use when Cargo.toml selects Bevy.