Bell Boundaries and Orthogonal Reconstruction Geometry Under Declared Observer-Level Channels

This internally audited release presents the corrected TEST03 Phase 3 Bell reconstruction programme, comprising Tests 3B-3E. The studies use the validated TEST02 nonfactorizable amplitude construction as a frozen observer-level baseline and apply explicitly declared quantum channels to distinguish four independently calculated quantities: CHSH violation, quantum-state entanglement, correlation-basis geometry, and finite observer recoverability. TEST 3B separates the isotropic CHSH boundary, q = 1/sqrt(2), from the quantum-state entanglement boundary, q = 1/3. TEST 3C demonstrates channel-specific Bell and entanglement persistence under selective phase damping. TEST 3D validates blind recovery of an injected anisotropic correlation basis and identifies measurement-plane-dependent CHSH capacity. TEST 3E shows that correlation geometry can remain exactly present and finite-shot recoverable after CHSH loss, while factorized and degenerate controls prevent this persistence from being interpreted as unique evidence of entanglement, nonfactorizability, or phase ontology. All four studies and all 12 automated regression tests pass. The release classification is: CONDITIONAL TEST03 OBSERVER-LEVEL RECONSTRUCTION PASS The result establishes controlled mathematical consequences of declared observer-level channels acting on the TEST02 Bell reconstruction. It does not independently derive the TEST02 half-angle map, antisymmetric amplitude relation, or squared-projection rule from primitive phase dynamics. It also does not establish a unique phase cause, a microscopic environmental decoherence law, ontological generation of locality, or empirical preference over standard quantum mechanics. The deposit includes the complete reproducibility package, frozen architecture, executable source, automated tests, exact numerical outputs, finite-shot recovery outputs, figures, integrity manifests, and the accompanying technical report.