AZ Tools MCP - Real-time Azure Infrastructure Querying
Secure Azure Resource Management and AI Querying
Developed Azure inventory and operations MCP capabilities (AzTools + AzOperator) that combine multi-subscription discovery, IAM visibility, backup and monitoring signals, Greenbone VM target sync, and real-time Azure API retrieval for AI-assisted infrastructure analysis.
Client
Luxottica
Completion
2 months
Category
Secure AI & Automation
Also: Security & Compliance
Part of platform
KC Enterprise AI & Operations Platform →Situation
The organization needed the ability to query Azure infrastructure data in real time for AI-powered analysis, executive reporting, and infrastructure optimization. Traditional approaches required persistent storage of sensitive infrastructure data, creating security risks, while fragmented inventories made it harder to reason about IAM permissions, backup posture, and resource health across subscriptions.
Task
Design and implement a secure Azure resource management capability that retrieves real-time infrastructure data through device login and 2FA authentication, provides multi-subscription discovery and operational insight, and supports AI-assisted reporting without permanently storing sensitive infrastructure data.
Action
→Deployed AzTools and AzOperator on the shared production platform with Traefik routing for Streamlit UI and MCP server access
→Integrated Greenbone adapter API for VM host target sync from Azure inventory into vulnerability management workflows
→Developed custom MCP tool for Azure resource querying using Python AZ CLI authentication model
→Expanded the Azure inventory surface into a broader resource management platform using Azure SDKs, Streamlit-style pages, and API services for multi-subscription discovery
→Implemented Microsoft Device login workflow with Multi-Factor Authentication (2FA) to retrieve secure bearer tokens
→Created real-time querying capability for Azure resources including subscriptions, resource groups, and Virtual Machines
→Modeled IAM permissions, backup management, monitoring signals, and infrastructure optimization views as public-safe operational overlays
→Successfully demonstrated retrieval of configuration data for 47 Virtual Machines across two Azure subscriptions
→Designed ephemeral data handling ensuring bearer tokens and associated infrastructure data are automatically erased upon conclusion of MCP discussions
→Integrated with Open WebUI (Lux GPT) to enable conversational queries of Azure infrastructure
→Developed executive reporting capabilities generating factually dense, real-time cybersecurity reports highlighting critical findings
→Implemented secure token management with automatic expiration and cleanup to prevent data persistence
Results
✓Enabled real-time, secure querying of Azure infrastructure resources without persistent data storage, significantly reducing security risks
✓Created a richer operational source of truth for Azure resources, permissions, backup posture, and monitoring signals across subscription boundaries
✓Demonstrated successful retrieval of comprehensive infrastructure data (47 VMs across 2 subscriptions) for analysis and reporting
✓Generated factually dense executive reports with real-time data, such as identifying outdated Red Hat versions across infrastructure
✓Provided seamless integration with conversational AI interfaces, enabling natural language queries of Azure resources
✓Established secure authentication workflow using device login and 2FA, meeting enterprise security requirements
✓Eliminated security concerns around persistent storage of sensitive infrastructure data through ephemeral data handling
✓Enabled faster decision-making through real-time infrastructure insights and automated executive reporting
✓Created foundation for AI-powered infrastructure management and analysis workflows