Game Asset Generator logo

Game Asset Generator

Community
notque
game-asset-generator

Deterministic palette/matrix pixel art (not AI). Use for procedural tile art, color-quantized output, matrix sprites.

Overview

Publishernotque
Repositoryvexjoy-agent
Skill namegame-asset-generator
Stars
443
Forks
50
Bundled files
8
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 8 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by notque on GitHub. Read the source before you install it.

Installation

Install the Game Asset Generator AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/notque/vexjoy-agent/tree/main/skills/game/game-asset-generator
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/notque/vexjoy-agent.git /tmp/vexjoy-agent
mkdir -p .claude/skills
cp -r /tmp/vexjoy-agent/skills/game/game-asset-generator .claude/skills/game-asset-generator
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Game Asset Generator in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Game Asset Generator on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Game Asset Generator is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Game Asset Generator Skill

Overview

This skill generates game-ready assets (3D models, Gaussian Splat environments, 2D sprites, images/textures) using AI APIs and free asset sources. It follows a three-phase workflow: DETECT the asset type -> GENERATE via the appropriate API or source -> INTEGRATE into the game. Only the relevant reference is loaded per task -- do not load all references upfront.

Scope: Use for AI-generated 3D models, world environments, pixel art sprites, concept art, textures, and sourcing free pre-built assets. Keep game engine scripting, physics, game loop logic, and shader authoring in threejs-builder after asset generation.


Phase 1: DETECT

Goal: Identify the asset type from the request and load the single corresponding reference.

Step 1: Classify the request

Signal in requestAsset TypeReference to load
"3D model", "character model", "generate model", GLB, mesh, rig, animate, humanoid3D Modelreferences/meshyai.md
"environment", "world", "scene background", "gaussian splat", "splat", volumetricEnvironmentreferences/worldlabs.md
"sprite", "pixel art", "2D character", "tile", "tileset", canvas sprite2D Spritereferences/pixel-art-sprites.md
"image", "texture", "concept art", "icon", "generate image", chroma keyImage / Texturereferences/fal-ai-image.md
No API key available, "free asset", "find model", "download asset", generation failedExisting Assetsreferences/asset-sources.md

If the request is ambiguous between 3D Model and Image, ask: "Do you need a 3D mesh (GLB file for a Three.js scene) or a 2D image/texture?"

Step 2: Check API key availability

Before calling any paid API, verify the required key exists:

bash
grep -E "MESHY_API_KEY|WLT_API_KEY|FAL_KEY" ~/.env 2>/dev/null

If the required key is missing, the fallback chain applies -- load references/asset-sources.md alongside the primary reference.

Fallback chain: Meshy API -> Sketchfab search -> Poly Haven -> Poly.pizza -> BoxGeometry placeholder. All sources output GLB into the same path so game loading code does not change.

Gate: Asset type identified, relevant reference loaded. Proceed to Phase 2 only when gate passes.


Phase 2: GENERATE

Goal: Call the API or source to produce the asset. Follow the loaded reference exactly -- it is the authoritative guide for its API.

Core constraints (all asset types):

  • Download immediately -- Meshy retains assets only 3 days; World Labs SPZ files expire similarly. Never assume a URL will be valid tomorrow.
  • Output to a stable path -- write assets to public/assets/ or an equivalent game-accessible directory so integration code does not need path changes per asset.
  • Save the .meta.json sidecar -- every generated asset gets a .meta.json recording the prompt, model, generation timestamp, and asset ID. Required for regeneration and auditing.
  • Validate the output file -- after download, confirm file size > 0 and extension matches expected type (GLB, SPZ, PNG, etc.) before proceeding.

Per-type generation summary (read the reference for full API details):

3D Model (Meshy): Two-step pipeline -- preview (fast, low quality, confirms prompt works) -> refine (full quality). Auto-rig only for humanoids meeting all criteria: bipedal, textured, clearly defined limbs. Animate rigged models with walk/run/idle presets. Post-process with scripts/optimize-glb.mjs for 80-95% size reduction before integration.

Environment (World Labs): Upload reference image (preferred over text-only) -> poll 3-8 minutes -> download SPZ + GLB collider + panorama JPG. Y-axis flip (rotation.x = Math.PI) required after loading into Three.js scene.

2D Sprite (code-only): Canvas-based generation -- no API call. Load references/pixel-art-sprites.md and generate sprites from the palette and matrix system defined there. Works without any API key.

Image / Texture (fal.ai): Queue-based API -- submit job -> poll for result. Choose model endpoint based on need (GPT Image 1.5 for transparency, Nano Banana 2 for speed). Use #00FF00 chroma-key background when the asset needs transparency extraction.

Gate: Asset file downloaded and validated (size > 0, correct extension). .meta.json saved. Proceed to Phase 3 only when gate passes.


Phase 3: INTEGRATE

Goal: Load the generated asset into the game scene correctly.

Core constraint: Use SkeletonUtils.clone() -- never .clone() -- for animated models. Regular .clone() breaks skeleton bindings and leaves the model in a permanent T-pose. This is the single most common integration failure with rigged GLBs.

javascript
import { SkeletonUtils } from 'three/addons/utils/SkeletonUtils.js';

// Load once, clone for each instance
loader.load('/assets/character.glb', (gltf) => {
  const instance = SkeletonUtils.clone(gltf.scene);
  scene.add(instance);
});

GLB loading (Three.js):

javascript
import { GLTFLoader } from 'three/addons/loaders/GLTFLoader.js';
import { DRACOLoader } from 'three/addons/loaders/DRACOLoader.js';

const dracoLoader = new DRACOLoader();
dracoLoader.setDecoderPath('https://www.gstatic.com/draco/versioned/decoders/1.5.6/');

const loader = new GLTFLoader();
loader.setDRACOLoader(dracoLoader); // Required for Draco-compressed GLBs from Meshy optimizer
loader.load('/assets/model.glb', (gltf) => {
  const model = gltf.scene;
  const box = new THREE.Box3().setFromObject(model);
  const center = box.getCenter(new THREE.Vector3());
  model.position.sub(center);
  scene.add(model);
});

Gaussian Splat (World Labs) -- see references/worldlabs.md for the @sparkjsdev/spark SplatMesh integration. The Y-axis flip and raycast direction inversion are required, not optional.

Animation playback:

javascript
const mixer = new THREE.AnimationMixer(model);
const action = mixer.clipAction(gltf.animations[0]); // walk/run/idle from Meshy
action.play();

// In animation loop:
mixer.update(deltaTime);

Gate: Asset visible in scene. No console errors. Animations play if applicable.


Error Handling

Error: "GLB loads but model is in T-pose"

Cause: Used .clone() instead of SkeletonUtils.clone() on a rigged model. Solution: Replace gltf.scene.clone() with SkeletonUtils.clone(gltf.scene). Import from three/addons/utils/SkeletonUtils.js.

Error: "Meshy task stuck in PENDING"

Cause: API key invalid, quota exceeded, or Meshy service issue. Solution:

  1. Verify MESHY_API_KEY is set in ~/.env and the value is current
  2. Check quota at app.meshy.ai
  3. If quota exhausted, fall through to references/asset-sources.md fallback chain
  4. Use status mode: node scripts/meshy-generate.mjs status <task_id>

Error: "Downloaded GLB is 0 bytes or corrupt"

Cause: URL expired (Meshy 3-day limit) or network error during download. Solution: Regenerate -- do not attempt to repair a corrupt GLB. Resubmit using the prompt saved in .meta.json.

Error: "Gaussian Splat renders but objects fall through floor"

Cause: Raycast direction inverted after Y-axis flip on the SplatMesh. Solution: Load references/worldlabs.md -- the fix is in the raycast inversion section.

Error: "fal.ai request returns 401"

Cause: FAL_KEY missing or incorrectly formatted. fal.ai uses Key $FAL_KEY format (not Bearer). Solution: Confirm FAL_KEY is in ~/.env. Authorization header must be Key <your-key> -- not Bearer <your-key>.

Error: "gltf-transform command not found"

Cause: @gltf-transform/cli not installed globally. Solution: npm install -g @gltf-transform/cli


Deep References

Load the reference matching the asset type from Phase 1 DETECT.

Asset typeReference
3D model (Meshy API)references/meshyai.md
Environment (World Labs)references/worldlabs.md
Image / texture (fal.ai)references/fal-ai-image.md
Free assets (Sketchfab, Poly Haven)references/asset-sources.md
2D sprite (canvas, no API)references/pixel-art-sprites.md

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Game Asset Generator AI skill do?

Deterministic palette/matrix pixel art (not AI). Use for procedural tile art, color-quantized output, matrix sprites.

Why use Game Asset Generator on TypingMind?

Because you install it once and use it with any model. Game Asset Generator is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Game Asset Generator in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/notque/vexjoy-agent/tree/main/skills/game/game-asset-generator. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Game Asset Generator?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Game Asset Generator?

As many as you like. As long as a model supports skills, you can use Game Asset Generator with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Game Asset Generator AI skill free?

Yes. It is published on GitHub by notque under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇