The "Agentic Organisation": Governing Multi-Agent Developer Teams with Nutanix Agent Gateway and MCP
summary
Central control plane enforces granular tool filtering and least-privilege access for multi-agent workflows using Nutanix Agent Gateway and Model Context Protocol (MCP), mapping agents to isolated Kubernetes namespaces and restricting which models and databases each agent can call. Built-in logging, token and rate limits, and an on‑prem local MCP server support air‑gapped deployments; layered fallback (local Llama → cloud models like OpenAI/Gemini) provides resiliency. Identity integrates with Azure/LDAP and policies block unrestricted agent access.
AI-written summary · about
chapters (10)
- 0:00Introduction to Nutanix
- 1:09Multi-agent development teams
- 3:47Autonomous agents collaboration
- 5:53Security risks of uncontrolled agents
- 7:21Nutanix Agent Gateway solution
- 10:17Central policy control features
- 12:27Fallback and resiliency models
- 13:33Local MCP server benefits
- 15:01Granular access and zero trust
- 16:28Q&A session
related talks
26mArchitecting Enterprise Agents Ecosystem: Standardizing APIs with MCP and Orch…
25mWho Let the Agents Out? Securing AI Workflows the Right Way
44mMCP Security: Keep Your AI Agents from Spilling the Tea
36mPutting AI Agents to Work: A New Era of Open Source Connectivity with MCP
08magentctl: A CLI for Managing AI Agents & MCP Servers Like Infra
40mBuilding Agents with Agent Bricks and MCP
40mIn AI We Trust? Securing the Future, One Agent at a Time
28mAgentic Development at Enterprise Scale: RAG, MCP & Guardrails