ARK-447 — ExecutionProof Authorization Boundary Experiment — PASS

Verdict: PASS (strong). Noise-mitigation comparison — Pauli twirling vs. unprotected baseline — for the authorization boundary on IBM Quantum ibm_marrakesh (Heron r2), Q_A=1 / Q_P=2. Both configurations passed all quantitative criteria using raw DENY leakage, and Pauli twirling showed a modest, statistically significant improvement in ALLOW fidelity over the unprotected baseline. Scope: this compares Pauli twirling to baseline only — Dynamical Decoupling was omitted due to scheduling complexity, so this is not a complete noise-mitigation survey. Boundary characterization on one device, not a general security claim. 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.