Threejs Gameplay Systems

Three.js Gameplay Systems

Create a playable, maintainable loop with clear ownership and responsive controls.

Read By Need

• Always for implementation: references/gameplay-workflows.md.
• New games, major loop changes, levels, encounters, progression, or difficulty:

references/game-design-level-design.md.

• Physics or collision-heavy work: references/physics-engine-selection.md.
• Model axes, constrained motion, multiple moving actors, or terrain queries:

../../references/spatial-contracts.md.

• Feel, impact, juice, or polished gameplay: references/game-feel.md.
• Before claiming broad gameplay complete: references/quality-gates.md.
• Generated audio or 3D gameplay assets:

../../references/mint-mcp-assets.md.

Workflow

1. Inspect scripts, dependencies, loop ownership, input, camera, state, entities,

UI, diagnostics, and current gameplay.

1. Define the player promise, primary verb, objective, pressure, reward,

fail/retry path, skill expression, and non-goals.

1. Plan the first playable space or encounter: start, decision, threat, reward,

escalation, recovery, and landmarks.

1. Choose small ownership boundaries such as core, game, entities,

systems, assets, ui, and tests.

1. Implement in playable increments: input, state, entity, collision/physics,

feedback, HUD/audio event, and diagnostics.

1. Tune movement, camera, timing, difficulty, feedback, and restart through

short play loops.

1. Verify build, browser rendering, real input, objective progress, and

fail/retry when relevant.

Gameplay code emits audio events. Mint-produced audio files belong in a stable runtime asset matrix; Mint MCP calls never belong in browser code.

New-Game Scaffold

python3 /scripts/create_threejs_game.py ./my-game

Use --force only when overwriting the target is intended.

Technical Defaults

• TypeScript, Vite, and Three.js modules.
• Custom collision for simple arcade triggers; Rapier for robust rigid-body

gameplay; cannon-es only when its smaller JavaScript footprint is useful.

• Fixed-step simulation for physics and seeded randomness for deterministic QA.
• Shared resources, explicit update order, and allocation-light hot paths.
• Web Audio for playback; Mint MCP for generated production audio.

Final Response

Report player-facing behavior, controls, design/level decisions when relevant, architecture, tuned values, verification evidence, assets, and remaining edge cases.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: mintdotgg
• Source: mintdotgg/mint-threejs-skills
• License: MIT
• Homepage: https://mcp.mint.gg

Install and usage instructions live in the source repository linked above.