GAME skill. Matt Shumer's aim prompt as a slash command — build a game against a named reference. Fill the three paragraphs. Run them. Stop when you stop it.
No harness. No state machine. No helper scripts. Pure prompt.
Works on Claude Code (/gauntlet-loop, /loop) and Codex ($gauntlet-loop, /goal).
/gauntlet-loop a COD-style FPS in ThreeJS
Built on Matt Shumer (@mattshumer_) cooking the 152-word Call of Duty prompt.
- Original prompt + demo: github.com/mshumer/Claude-of-Duty
- This skill: same prompt, filled and executed on invoke — nothing bolted on
He cooked the aim. This just makes it one command.
git clone https://github.com/duolahypercho/gauntlet-loop.git
cd gauntlet-loop
./install.shInstalls into ~/.claude/skills/gauntlet-loop + /gauntlet-loop, and symlinks Codex to the same skill.
/gauntlet-loop a COD-style FPS in ThreeJS
/gauntlet-loop marketing site against Linear in Next.js
$gauntlet-loop Hades-like roguelike in Godot
The agent fills this shape and runs it (Claude verbs shown):
I want you to build [THING] at the level of [REFERENCE]. It should
be utterly perfect, [LOOK], with every single thing done at
[TIER] quality, from [AREA_1] to [AREA_2] to anything you could think of.
Fan out sub-agents and have sub-agents tackle each one individually so that the [THING]
is utterly perfect. You should /loop on each item and have a separate sub-agent check it
[CHECK] to ensure it is [TIER]. That separate sub-agent should
be a really harsh critic, and if it isn't [TIER], it should keep going.
Don't stop until each sub-agent is utterly wowed with the quality when compared with
[REFERENCE]. It should literally compare them side by side blind and say which
one looks better. Do this in [STACK]. /loop until it's utterly perfect.
Fan out sub-agents and ultracode.
Say "compose only" if you just want the filled text without the run.
You are the brake. The loop will not finish on its own.
gauntlet-loop/
├── README.md
├── LICENSE
├── install.sh
├── commands/gauntlet-loop.md
└── skills/gauntlet-loop/
├── SKILL.md # boot
├── AGENTS.md # the prompt + what not to invent
├── CLAUDE.md # /loop + ultracode
├── CODEX.md # /goal
├── examples.md # filled game prompts
├── ASSETS.md # image gen vs Blender
└── BLENDER_MCP.md # Blender MCP setup + use
MIT