An open method for teachers
Skill trees your students can actually see.
A teacher-written map of a subject, plus instructions that turn any AI chat into a practice partner — one that holds back what it could give away.
What this is
The graph is the interface, not the engine.
Prerequisite graphs are not new — adaptive systems have computed them for decades. What they do not do is show them. Here the graph is the interface: the student sees the whole structure of a subject, chooses where to start, and can tell what any one skill rests on.
Nothing here calls a language model. The tool composes a teacher-written instruction that you paste into whichever AI chat you already use. No vendor lock-in, no cost per student, no data processing agreement — and every instruction is written to make the model hold something back, so the student still does the thinking.
Start here
Existing skill trees
Browse the subjects already decomposed — what a finished tree looks like, and what a node actually contains.
02Make your own
The decomposition specification, the instructions, and a template for starting a new subject — in your language, for your syllabus.
03Research
The cognitive grounding behind the design, and the paper describing why a visible, teacher-written graph is the point.