This is my 4th newsletter on AI and .NET topics š focussing on AI use cases, software architecture, and production-grade patterns.
If youāre working hands-on with .NET, AI, or relevant enterprise architectures, feel free to connect š, follow š, or jump into the discussion.
AI Toolkit for VS Code: Microsoft released AI Toolkit 0.30.0, which materially advances VS Codeās support for agent development by improving discoverability, structured evaluation, real debugging, and deeper integration with Microsoft Foundry workflows. These changes reduce friction for local development and early testing of agent orchestration before moving to CI/CD pipelines, signaling that the toolkit is approaching productionāoriented usability, not just experimental play. The enhancements also underscore a shift toward composable agent workflows where debugging and observability are firstāclass concerns.
GitHub Copilot Testing for .NET: The GitHub Copilot Testing for .NET feature has reached general availability in Visual Studio 2026 (18.3), enabling developers to generate, build, and run unit tests with AI assistance directly in the IDE. This capability interprets highālevel intent (via @Test prompts) and automatically generates structured test projects, executes tests, and iterates to stable outcomes. For architects, this elevates discussions about AIāassisted development workflows into mainstream tooling, but also necessitates clear guardrails around test quality, determinism, and integration with CI pipelines.
šhttps://devblogs.microsoft.com/dotnet/github-copilot-testing-for-dotnet-available-in-visual-studio/
VS Code 1.109 Enhances Agent Session Support: The January release of Visual Studio Code 1.109 introduces key improvements to GitHub Copilot agent workflows, including richer multiāagent session management, extensibility for agent models and tools, and performance/UX optimizations. This release reflects significant maturation of ināeditor multiāagent orchestration, which encourages architects to think about how agent lifecycles, state management, and security contexts are tracked and instrumented during development and experimentation.