Azure MCP Server

Instructions

Implementation Reference

• mcp_server_azure/azure_server.py:219-225 (handler)

  Executes the blob_read tool by creating a blob client, downloading the blob content, decoding it as UTF-8, and returning it as TextContent.

  elif name == "blob_read":
      blob_client = blob_service_client.get_blob_client(
          container=arguments["container_name"], blob=arguments["blob_name"]
      )
      downloader = blob_client.download_blob()
      content = downloader.readall().decode("utf-8")
      return [TextContent(type="text", text=content)]

• mcp_server_azure/azure_tools.py:94-111 (schema)

  Defines the schema for the blob_read tool, including input parameters container_name and blob_name as required strings.

  Tool(
      name="blob_read",
      description="Read a blob's content from Blob Storage",
      inputSchema={
          "type": "object",
          "properties": {
              "container_name": {
                  "type": "string",
                  "description": "Name of the Blob Storage container",
              },
              "blob_name": {
                  "type": "string",
                  "description": "Name of the blob to read",
              },
          },
          "required": ["container_name", "blob_name"],
      },
  ),

• mcp_server_azure/azure_server.py:172-176 (registration)

  Registers the blob_read tool (among others) by returning the list of Azure tools from get_azure_tools() in the list_tools handler.

  async def list_tools() -> list[Tool]:
      """List available Azure tools"""
      logger.debug("Handling list_tools request")
      return get_azure_tools()  # Use get_azure_tools