This environment is a research and testing lab exploring how AI agents, powered by Model Context Protocol (MCP) servers, can integrate with modern security infrastructure including Duo Security and Zero Trust Secure Access platforms.
MCP Servers AI Agents Duo Security Zero Trust Secure AccessMCP is an open standard developed by Anthropic that defines how AI models communicate with external tools, data sources, and services. It provides a unified interface so that AI assistants can securely connect to databases, APIs, file systems, and enterprise platforms — without requiring custom integrations for every tool.
An MCP server is a lightweight service that exposes capabilities (tools, resources, prompts) to an AI model. The AI agent calls the MCP server to read data, trigger actions, or query external systems. Examples include: a file system MCP server, a database MCP server, a security policy MCP server, or a Duo authentication MCP server.
An AI agent is an AI model that can autonomously plan, reason, and execute multi-step tasks by calling tools (via MCP servers). Unlike a simple chatbot, an agent can take actions: querying a security log, triggering a Duo push notification, evaluating access policy, or updating a configuration — all orchestrated by the AI model itself.
Exploring how AI agents can interact with Duo Security for adaptive MFA, risk-based authentication decisions, and automated policy enforcement.
Testing AI-driven access control where agents evaluate user context, device posture, and behaviour patterns to grant or deny access in real time.
Building and connecting MCP servers that expose security tools and enterprise APIs to AI models in a controlled, auditable way.
Researching how AI agents can assist in threat detection, incident response, and compliance reporting within a secure access framework.
The MCP Server is live. Query Azure resources, manage tools, and monitor sessions.
MCP Server Manager View DocumentsWebserver secured with a trusted certificate via Azure DNS automation.
Public A record created and confirmed resolving correctly.
Deploy first MCP servers exposing security tools to AI agents.
Connect AI agents to Duo Security APIs via a dedicated MCP server.
Implement AI-driven access policies and adaptive security controls.