No token launch. No deployment. No market simulation.
Just:
– explicit state surface – deterministic events – invariant layers (correctness + safety) – stress scenarios (S01–S22)
The goal is structural falsifiability.
If the invariant system collapses under adversarial sequencing, I want to know.
Blueprint: https://github.com/ArturGrandi/gt2-research-track/blob/main/...