R&D-001
HORIZIN Agent Command Layer
A HORIZIN-owned control layer for enterprise AI agents that keeps identity, permissions, human approvals, audit evidence and runtime governance outside any single model provider.
Multi-runtime
Custom code, OpenAI integration boundaries, Microsoft Agent 365 integration path and MCP security controls behind a common invocation model.
Authenticated prototype
Current v0.7 service prototype has 49 automated tests passing across the command layer, runtime conformance and authenticated service controls.
Prototype status
This is controlled internal R&D and is not represented as production-ready. Live provider credentials and production infrastructure remain separate deployment stages.