Explaining how something works has always cost more than understanding it, which is why most technical teaching still leans on paragraphs and boxes with arrows when a moving picture of the mechanism would do the job far better.
Anatomy, an open-source Claude Code skill from Ryan (@wheresryan22), brings that cost down to a single line. You describe a machine or an idea, and the skill invents a physical object whose working is the concept, then builds it as an isometric figure you can rotate, take apart and inspect, with WebGL shaders for fire and water and individual parts that actually move. In the repo a cache key becomes a pin-tumbler lock that only turns when every pin matches, and every number in the live readout comes from the same small model that drives the parts.
It is MIT-licensed, needs nothing beyond Node’s built-in modules, and also runs in OpenCode and Codex, and Meng To has already used it to explain how Seedream 5.0 splits an image into layers. Guillermo Rauch called it the future of learning, and the economics back him up, because an explainer that used to need a designer and a week now needs a clear prompt and a patient session, which puts the mechanism within reach of every engineer writing docs, every teacher preparing a class and every team onboarding someone new.