Bridging LLM Agents and Data Spaces with MCP-Based Architectural Mediation

Summary

This article addresses the difficulty of integrating probabilistic LLM agents with data spaces, which support sovereign and policy-governed data sharing across organizational boundaries. It proposes an architectural mediation layer based on the Model Context Protocol (MCP), implemented through the Eunomia Agent. The layer translates data-space capabilities into structured, schema-driven tools that agents can discover and invoke while retaining the underlying governance constraints. A prototype demonstrates end-to-end interaction across catalog discovery, metadata retrieval, and data-service invocation. The demonstration does not require changes to existing data-space components. The reported results indicate that protocol-based mediation can provide an interoperable, standards-aligned way to integrate agents into data-space ecosystems. The article presents the approach as practical guidance for organizations introducing AI automation into governed data-sharing environments, with compliance, interoperability, and separation of architectural responsibilities as key objectives.