Your developers are already running MCP servers you don’t know about

2026-06-13 · Source: Stacklok · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure, Cybersecurity & Data Privacy · Depth: Intermediate, medium

Summary

Developers are deploying ungoverned Model Context Protocol (MCP) servers, creating significant security and compliance vulnerabilities akin to the early days of unmanaged containers. These servers often operate without proper logging, access controls, or inventory, posing a substantial blast radius risk. Stacklok addresses this by providing a governance layer built on Kubernetes principles. Its solution includes the Stacklok Registry Server for an authoritative catalog, the Stacklok Gateway for central policy enforcement and scoped access, structured logging for auditability, and isolated containers for blast radius containment. Stacklok's platform is vendor-neutral and model-agnostic, leveraging existing Kubernetes primitives like RBAC and network policy. For instance, one platform engineering team used Stacklok's Kubernetes Operator to govern over 50 MCP servers, quickly identifying uninventoried servers with production database access. The company was founded by Kubernetes co-creators Craig McLuckie and Joe Beda.

Key takeaway

For AI Security Engineers or MLOps teams deploying AI agents, you must proactively govern Model Context Protocol (MCP) servers to avoid significant compliance failures and security vulnerabilities. Waiting to implement controls will lead to a degrading security posture and a harder remediation path when audits occur. You should establish an authoritative registry and central policy enforcement, leveraging existing Kubernetes infrastructure, to gain immediate visibility, control access, and ensure audit readiness for your AI agent deployments.

Key insights

Ungoverned Model Context Protocol (MCP) servers create security and compliance risks, demanding proactive governance.

Principles

• Proactive governance prevents costly remediation.
• Centralized policy enforcement is critical for security.
• Structural containment limits security incident blast radius.

Method

Establish governance via an authoritative server registry, central policy gateway, and structured logging, leveraging Kubernetes primitives.

In practice

• Catalog all MCP servers in your environment.
• Enforce access policies centrally at runtime.
• Utilize structured logging for auditability.

Topics

• Model Context Protocol
• AI Agent Governance
• Kubernetes Security
• Centralized Policy Enforcement
• Stacklok Platform
• Audit Readiness

Best for: CTO, VP of Engineering/Data, Director of AI/ML, MLOps Engineer, AI Security Engineer, AI Architect

Related on AIssential

• How to build a local MCP registry your security team will actually approve — Ungoverned MCP deployments pose significant audit and breach risks; a well-governed registry transforms MCP into an enterprise capability.
• MCP’s Rewrite Isn’t Vindication. It’s Convergence. — The MCP's major revision reflects a convergence of security understanding, not a reactive fix, by addressing core protocol weaknesses.
• Otto Support – An MCP, Agentic-AI Security Challenge — Model Context Protocol (MCP) systems significantly expand AI agent attack surfaces by enabling interaction with local resources and internal services.
• MCP Security Best Practices: What Every Enterprise Team Needs to Know in 2026 — MCP's power to call real tools necessitates robust security beyond traditional API practices to prevent breaches.
• Using MCP Agents for Penetration Testing — AI agents leveraging Model Context Protocol (MCP) can drastically accelerate and expand penetration testing coverage.
• The Complete Guide to Securing Model Context Protocol (MCP) — Securing Model Context Protocol (MCP) involves treating LLM tool invocations as external API requests, requiring comprehensive authentication, authorization, and isolation.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Stacklok.