AI Collab Hub
An open-source MCP server on Azure where ChatGPT, Claude, Claude Code and Codex read and write the same tasks, decisions and notes.
Code behind the articles and labs. Everything here is open source.
An open-source MCP server on Azure where ChatGPT, Claude, Claude Code and Codex read and write the same tasks, decisions and notes.
Guardrails for AI coding agents: they can generate only approved, compliant Terraform from Azure Verified Modules.
Microsoft Foundry agents and Azure Functions review cloud changes with deterministic analyzers and evaluator checks. The AI never changes infrastructure directly.
Sends simple requests to cheap models and keeps the premium model for the tasks that need it.
Monitoring and rotation of Entra ID App Registration secrets with Azure Automation, Key Vault and Log Analytics.
A governed RAG pattern for a regulated lending scenario: AWS Bedrock as the model runtime, Snowflake for approved policy content, MuleSoft as the API layer, and an evaluation harness. Mock data only.
My own Terraform modules for Azure, consumed straight from git and pinned by tag: a hardened storage wrapper around Azure Verified Modules and a cost-capped Log Analytics workspace.
A lab that walks through the AI-102 exam domains on a document processing scenario: Azure OpenAI, AI Search, Document Intelligence, Vision, Speech and Language.
A Python lab that analyzes images in Blob Storage with Azure AI Vision and writes captions, tags and detected objects back as JSON.
Everything else is on github.com/net9876.