ARK-448 — ExecutionProof Authorization Boundary Experiment — GATE-STOP (kill-gate not satisfied)

Outcome: GATE-STOP — aborted at the SPAM gate (honest, preregistered stop). No verdict is claimed. Intended as a Dynamical Decoupling vs. baseline comparison on IBM Quantum ibm_marrakesh (Heron r2), Q_A=1 / Q_P=2. The experiment was locked, qubits selected, and the preregistered SPAM gate executed on real hardware. The SPAM gate FAILED: payload readout symmetry deviated 0.02197 from the ideal 0.5 (threshold ±0.02), ≈ 2.0σ at 2,048 shots — consistent with a marginal fluctuation / minor readout drift. Per the LOCKED preregistration, the principal 4-arm job was NOT submitted and execution halted. No DD-vs-baseline data was obtained. This record is preserved as a preregistered abort demonstrating fail-closed discipline, not a result about dynamical decoupling. Series: Part of the ExecutionProof™ ARK (Authorization Reality Kernel) series by Remnant Fieldworks Inc. — a preregistered falsification program probing verify-then-execute authorization boundaries on IBM Quantum hardware and classical testbeds. Scope & limitations: Classical/software boundary-logic testbed or quantum hardware experiment. This is an experimental testbed, not a production security claim. Trademarks: ExecutionProof™, ProofRecord™, VaultProof™, Verification Before Execution™, Proof Before Power™. Remnant Fieldworks Inc.