Thomas Thornton is a cloud and DevOps specialist focused on Microsoft Azure, Azure DevOps, GitHub, and Terraform. Through practical tutorials and real-world troubleshooting, he helps developers and platform teams build, secure, and automate modern cloud solutions.
28 articles from this blog
Explains why agent skill descriptions are critical for AI coding agents to select relevant skills, improving consistency and reducing prompt repetition.
A quick demo of using HolmesGPT with Azure AI Foundry to troubleshoot a Kubernetes pod stuck in pending state.
Guide to scanning AI skill repositories for security risks using NVIDIA SkillSpector integrated with GitHub Actions CI/CD.
Explores necessary platform controls for scaling AI-assisted engineering, focusing on cost visibility, ownership, and governance.
Analysis of AI-assisted engineering evolving from productivity feature to platform capability, focusing on repeatability, safety, and cost.
Adding optional animated traffic flows to Draw.io diagrams using GitHub Copilot Agent skills to highlight specific paths.
Advice on keeping GitHub Copilot Agent Skills small and focused for better AI performance and predictable output.
Explores the challenges of Terraform module design and introduces a custom agent skill to address design decisions before code generation.
GitHub Copilot Rubber Duck uses a second AI model to review code plans, catching subtle errors in multi-file tasks.
Analyzes design choices for effective GitHub Copilot Agent skills, focusing on task shape, descriptions, boundaries, and testing.
Explains how Agent Package Manager (APM) solves packaging and versioning problems for GitHub Copilot agents and skills across teams.
Using Excalidraw with MCP and GitHub Copilot to create architecture diagrams directly in VS Code, keeping them part of the design workflow.
Integrates live Azure pricing into GitHub Copilot via an MCP skill, enabling cost estimation during architecture design without switching tools.
Extends an AI diagramming skill to generate AWS architecture diagrams using Draw.io MCP, improving on generic tools for clarity and reliability.
Automating Terraform provider upgrades using GitHub Agentic Workflows and MCP servers to generate safe pull requests.
Introducing an Azure Diagram Agent Skill that uses Draw.io MCP and GitHub Copilot to automate the generation of consistent, readable Azure architecture diagrams.
Explains how Draw.io MCP automates diagram generation from code, integrating diagrams into engineering workflows and version control.
A guide to integrating the GitHub Copilot SDK into GitHub Actions workflows to create an automated AI-powered Pull Request review assistant.
Explores how AI is transforming Infrastructure as Code (IaC) by enabling standardization, shifting from writing code to describing intent, and automating tasks.
Guide to identifying and disabling the deprecated 'trusted service connectivity' feature in Azure API Management before its March 2026 retirement.
