DTPP-Ω: A Decentralized Trust Propagation Protocol for Autonomous Agent Ecosystems — Anticipating Civilizational-Scale Trust Failures, 2026–2100

Autonomous agent ecosystems operating at civilisational scale require a trust layer that does not exist in contemporary agentic stacks. Identity protocols establish who 
an agent is. Permission systems gate what an agent may execute. Neither addresses how trust propagates between agents across a decentralised mesh without a central authority. This paper introduces DTPP-Ω (Decentralized Trust Propagation Protocol, Omega Architecture), a protocol designed to operate across the period 2026–2100. DTPP-Ω is built on eleven constitutional invariants, nine architectural components, a two-phase authorization engine, a bounded trust budget system, an influence graph layer, and a fully specified sentinel governance mechanism. The protocol's core design objective is trustworthy operation under uncertainty — not maximum trust, but bounded, explainable, auditable, locally computed, influence-aware trust that degrades gracefully under adversarial conditions. The architecture is motivated by backcasting analysis identifying the absence of an inter-agent trust propagation layer as a probable cause of hypothesised civilisational-scale alignment failures between 2039 and 2042.