Azure Self Help MCP Server

The Azure Self‑Help MCP Server turns Microsoft’s Guided Troubleshooter API into a first‑class tool for AI assistants. By exposing the Self‑Help endpoints as MCP tools, it lets a language model or semantic kernel agent create, inspect, advance, and terminate diagnostic sessions on any Azure resource. This bridges the gap between raw cloud APIs and conversational troubleshooting, enabling agents to ask a user for a resource URI, spin up a session, walk through each diagnostic step, and propose fixes—all without leaving the chat.

For developers building intelligent cloud‑management assistants, this server removes the need to write custom Azure SDK wrappers for every troubleshooting scenario. Instead of manually handling authentication, session state, and API payloads, the MCP server packages those concerns into a small set of declarative methods. The agent can simply call , then repeatedly invoke and as the user responds, mimicking the guided flow that a human would see in Azure Portal. The server also supports restarting or ending sessions, giving the user full control over the diagnostic lifecycle.

Key capabilities include:

• Session orchestration: The server manages Azure resource identifiers, solution IDs, and the lifecycle of a Self‑Help troubleshooter.
• Stateful step retrieval: Each call returns the current step, allowing the assistant to present clear next‑action prompts.
• Interactive continuation: By sending user input back to the API, the assistant can progress through complex troubleshooting trees.
• Authentication abstraction: Leveraging Azure’s DefaultAzureCredential, the server automatically scopes permissions based on the running context.

Typical use cases span from a DevOps bot that automatically diagnoses VM connectivity issues to an IT help‑desk assistant that walks users through resolving storage account errors. In production, the server can be paired with a Claude Desktop host or any MCP‑compatible application, enabling seamless integration into existing AI workflows without bespoke code. The result is a powerful, low‑maintenance diagnostic companion that scales with your cloud infrastructure and reduces mean time to resolution.