# New mediation layer lets LLM agents use governed data spaces

> Researchers released a prototype called the Eunomia Agent that connects large language model tools to data‑space services while keeping policy controls intact.

Oossa · 2026-09-28 · https://oossa.com/en/new-mediation-layer-lets-llm-agents-use-governed-data-spaces

A paper posted on arXiv on Sep 28, 2026 describes an architectural mediation approach named the Model Context Protocol (MCP). The authors built the Eunomia Agent to translate data‑space capabilities into structured tools that LLM agents can discover and call. The prototype shows end‑to‑end work for catalog search, metadata pull and data service calls without changing any existing data‑space components. The work aims to let companies add AI‑driven automation to regulated data‑sharing environments safely.

## The facts

- Paper posted Sep 28, 2026 on arXiv (arXiv:2609.30341) by Andres Munoz‑Arcentales.
- Prototype validates catalog discovery, metadata retrieval, and data service invocation via the Model Context Protocol and Eunomia Agent.

## Why it matters

It shows a practical way for organizations to let AI agents work with protected data without breaking governance rules.

## Sources & references

1. [Bridging LLM Agents and Data Spaces: An Architectural Mediation Approach using the Model Context Protocol](https://arxiv.org/abs/2609.30341) – arXiv, 2026-09-28

Last updated: 2026-09-28