Walk a feature, decision, claim or plan back through a person's own chain of first principles (kept as a framework file whose stages run from the most basic to the change it should make in the world), saying at each stage how the thing shows up there with evidence, which stage is its deepest root, whether it reaches a real change, and which stages it skips, then what to change so it does.
Nothing: your AI starts it with your words.
Connect the AI you already use to Trillion once. Then ask in your own words; your AI finds this workflow by what it does and runs it.
Say to your AI: “Use Trillion to walk a feature, decision, claim or plan back through a person's own chain of first principles (kept as a framework file whose stages run from the most basic to the change it should make in the world), saying at each stage how the thing shows up there with evidence, which stage is its deepest root, whether it reaches a real change, and which stages it skips, then what to change so it does”
What your AI calls: start_workflow {"workflow_id":"workflow:trace-to-first-principles"}