From traces to auditable decision trajectories.

The framework has three working layers: interaction provenance, loop provenance, and system provenance. None of these replaces infrastructure telemetry. They answer a different question: how did a Human–AI decision emerge?

The operational sequence is: ingest a trace, reconstruct a trajectory, inspect interventions and verification, then decide what can and cannot be audited. Metrics, when shown, are proposed indicators.