mcpField
T
three

three

github.com/magnus919Development15762 installsUpdated today
Repository

About three

three is a skill for building browser-based 3D scenes with Three.js. It is aimed at developers creating WebGL visualizations, configuring cameras and lighting, loading assets, and animating scenes, with guidance on cleanup and performance choices.

Key features

  1. 01Browser 3D scene setup
  2. 02Camera and lighting configuration
  3. 03Model asset loading
  4. 04Time-based animation loops
  5. 05Resource cleanup and performance guidance

Use cases

  1. 01Starting a Three.js scene from the basic template
  2. 02Creating WebGL visualizations
  3. 03Loading GLTF/GLB models and animating them

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/magnus919/agent-skills --skill three

After installing, type/threein Claude Code to confirm it loaded.

Bundled files

References and scripts installed along with SKILL.md.

README.md
evals/
evals.json
templates/
basic-scene.html

Details

From the project README

Three.js

Quick Start

Open templates/basic-scene.html in a modern browser or serve the repository over HTTP.

Use this skill for browser 3D scenes, WebGL visualization, camera and lighting setup, asset loading, or animation loops. Start with the smallest scene in templates/basic-scene.html, then add geometry, materials, lights, and controls only as required.

Core choices

  • Use PerspectiveCamera for natural 3D views and OrthographicCamera for technical or isometric views.
  • Keep animation time-based with Clock or elapsed timestamps.
  • Dispose geometries, materials, and textures when scenes are replaced.
  • Use GLTF/GLB for external models and keep assets local when reproducibility matters.
  • Prefer instancing and level of detail when object count becomes the bottleneck.

The template uses a CDN for a quick standalone demo. Pin a Three.js version for production.

Comments

Markdown supported