Maritime shipping carries over 80% of global trade by volume, yet the information systems underpinning this vast network remain fragmented, proprietary, and mutually distrustful. This paper presents the Cascading Visibility Model (CVM), a theoretical framework formalizing how a single upstream data failure propagates non-linearly through carrier, port, customs, warehouse, and trucking handoffs. We introduce the Entropy Amplification Index (EAI) as a normalized measure of information loss per handoff layer, with estimated values exceeding 0.6 at the carrier-to-port boundary and approaching 0.8 at port-to-customs. We further characterize the Multi-Layer Trust Deficit as a maritime-specific prisoner's dilemma in which rational data hoarding by individual actors produces collectively catastrophic coordination failures. To address these failures, we propose the Distributed Vessel Trust Pool (DVTP), a protocol-layer architecture enabling multi-party vessel verification without requiring raw data disclosure. The DVTP uses physical impossibility detection anchored to third-party-generated port event timestamps that vessels cannot falsify, combined with zero-knowledge proof logic to trigger automatic cascade holds across interconnected ports. Three adversarial scenarios are analyzed theoretically. We compare the DVTP with the Portbase model and TradeLens failure to derive governance lessons. We also present a formal research agenda of eight hypotheses for empirical validation through discrete-event simulation. This paper is a theoretical framework and research agenda contribution. The EAI estimates presented are model-derived under stated assumptions; the DVTP architecture and its adversarial analysis are theoretical proposals; and the simulation design is specified for future execution.