Anatomy - An AI Agent Skill That Turns Technical Concepts into Interactive 3D Diagrams

What happened

A developer shared Anatomy, a skill for AI coding agents. You ask the agent for a figure that explains a technical idea. Anatomy then invents a physical object that works the same way as the idea. The figure is interactive: it moves when you move your mouse or press keys. It is drawn in SVG, WebGL, and 3D.

How it works in plain words

Here is an example. A cache key is a label that helps a computer skip work it already did. Anatomy turns it into a pin-tumbler lock, which only opens when every pin lines up. The figure has many small, shaded parts in a framed card. It also shows a live readout. The numbers come from the same small model that moves the parts, so they match what you see.

Key facts

• Works with 5 agents: Claude Code, Antigravity, OpenCode, Codex, and MiniMax Code.
• Installs with one command: npx skills add wheresryan22/anatomy -g
• Needs Node.js version 20.10 or later.
• Has 0 dependencies, which means no extra software packages to install.
• Licensed under MIT, a free open-source license.

Why it matters

Many people find technical ideas hard to picture. A moving, hands-on model can make them easier to understand. Because the skill is free and works across several AI agents, students and developers can use it with the tools they already have.

What to watch

• Whether more AI agents add support for the skill.
• How well the figures match real systems as the examples grow.
• The skill's Limits section on the project page, which lists what it cannot do yet.