SPAR (Sovereign Physics Autonomous Review) - Claim-aware review framework for systems whose outputs can pass while their claims drift.

Lean 4 formalization of the neutral structural substrate for Structural Explainability, defining admissibility, separation, and substrate invariants independent of identity regimes.

This solver includes 2 versions: Classic (1 remaining peg in center hole), Max-Peg (play until no valid moves remained)

Formal constraint framework: C = muX.Phi(X). Five gradients, twelve laws, eight derivation theorems. Companion plain-language book: Wondering State.

Admissibility Rotation Corridor — closed constitutional runtime with visible interrupt geometry

Reference model for translating GRC controls into enforceable admissibility constraints for AI systems.

Pure Coulomb-Hooke constraint kernel — m x-double-dot + k x + F_c sign(x-dot) = u(t). Maps business parameters to oscillator dynamics. Only measures admissibility delta >= 0. No targets. No optimization.

Public learning kernel for consequence-aware AI governance: EXECUTE, TRANSFORM, DEFER, or REJECT before commitment.

Recursive admissibility architecture for non-terminal intelligence systems.

Minimal proof of execution-bound admissibility: no present-state proof at bind, no execution.

Constraint Alignment System — multi-domain constraint health engine mapping admissibility, structure, circulation, transfer, and pressure to a unified CAS state.

Public-safe PoC for admissibility-first execution governance in probabilistic and multi-agent AI systems.

Deterministic admissibility layer for interpretation proposals before verdict and execution.

Five-layer structural claim auditor. Asks: what breaks first, and was that break declared before movement? Outputs ADVANCE / REPAIR / HOLD / BLOCK.