Where today’s protocol stack stops — and where hybrid AI-native execution still lacks shared semantics.
Today’s stack already covers transport, tool access, agent collaboration, interface surfaces, identity foundations, observability, and packaging.
How tasks are constrained across local, edge, private, and public environments.
Where data, inference, and context may or may not go.
How privacy, latency, cost, sovereignty, and trust shape execution routing.
How systems reject, reroute, degrade, queue, or escalate execution under constraints.
How context moves across domains under controlled, auditable rules.
How execution evidence remains connected across systems and environments.
HAINP is intended to define the missing policy-aware execution layer for hybrid AI-native systems, rather than replace the protocols that already cover transport, tools, or agent communication.