Turns lessons and technical ideas into interactive 2D or 3D teaching figures that follow predict, try, observe, and explain, with offline bilingual demos.
teachable-figures: Predict-try-explain lessons
Builds interactive 2D or 3D teaching figures around one mechanism so learners predict, change a cause, observe the effect, and explain.
Source facts
- Repository
- ywedai/teachable-figures
- Last source activity
- November 3, 2026 at 02:30
- Detected SKILL.md language
- English
- Stars
- 305
- Forks
- 39
Examples
Author demos cover Transformer attention, chest anatomy, rocket structure, cone sections, a Ming robe, and gear ratios; bilingual offline demos at https://ywedai.github.io/teachable-figures (https://x.com/yiyibuilds/status/2116289366784168399).
Uses
Use for explorable learning graphics, mechanism simulations, cutaways, or spatial explainers—not for plain charts, dashboards, slide decks, static illustrations, or whole courses.
Prerequisites
A host that loads SKILL.md skills. Default guided mode asks at most five questions with defaults; direct mode skips when the user says so. Keep host lesson contracts and accessibility.
How to use
State the claim and misconception → choose mechanism kind and renderer (iso-figure, Anatomy craft, or custom when shape requires it) → build one figure with 2–4 controls and readouts → deliver standalone offline HTML or host-app output → list assumptions in the delivery note.
Limitations
A working figure is a teaching design, not learner-outcome evidence. Decline clinical anatomy, engineering simulation, CAD, or historical reconstruction presented as fact. One figure per mechanism by default.
Install options
The review-first prompt is selected by default. You can switch to a direct command or download a local copy.
Review the source files
Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.
Showing SKILL.md
| name | teachable-figures |
|---|---|
| description | Turn lessons, articles, chapters, or technical ideas into interactive teaching figures using predict → try → explain. Create reusable 2D widgets, working isometric SVG objects, mechanical assemblies and cutaways, exploded views, or true 3D turning models when spatial motion explains the concept. Use for explorable learning graphics, mechanism simulations, and mechanical or spatial explainers; retain the host app's lesson contracts and accessibility. Starts with a short guided interview unless the user asks to generate directly. Not for plain data charts, dashboards, slide decks, static illustrations, or whole courses. |