Azure AI Toolkit v0.28.1: Agent Skills Now Cost-Efficient for .NET Developers

TL;DR: Microsoft shipped major updates to the AI Toolkit for VS Code designed to streamline how you build, test, and deploy AI agents, focusing on aligning with the latest GitHub Copilot standards and enhancing support for enterprise-grade models via Microsoft Foundry. The headline: the transition from Copilot Instructions to Copilot Skills, which has equipped GitHub Copilot specialized skills on developing AI agents using Microsoft Foundry and Agent Framework in a cost-efficient way. If you’re shipping .NET agents on Azure, this is your signal to upgrade.

What Changed: Skills Over Instructions

In AI Toolkit, Microsoft migrated Copilot Tools from Custom Instructions to Agent Skills, which allows for a more capable integration within GitHub Copilot Chat. This might sound like nomenclature shuffling, but it’s a meaningful architectural shift. Agent Skills follow the open standard that Anthropic’s Model Context Protocol (MCP), a “USB-C for AI,” lets AI agents talk to external tools like databases, search engines, and APIs, and OpenAI and Microsoft have publicly embraced MCP. The practical upshot: your custom agent logic is no longer buried in IDE configuration. It’s now a first-class citizen in the Copilot ecosystem, meaning Copilot Chat can discover and invoke your agent’s specialized skills automatically.

Why This Matters for .NET Teams

Agent Framework is a production-ready, open-source framework that provides multi-agent orchestration and enterprise-grade features including built-in observability (OpenTelemetry), Microsoft Entra security integration, and responsible AI features including prompt injection protection and task adherence monitoring. With the AI Toolkit now aligned to Agent Skills, you can:

  1. Reduce boilerplate. The enhanced AIAgentExpert custom agent now has a deeper understanding of workflow code generation and evaluation planning/execution.
  2. Stay in VS Code. Profiling for Windows ML in version 0.28.0 introduces profiling features for Windows ML-based local models, allowing you to monitor performance and resource utilization directly within VS Code.
  3. Leverage Microsoft Foundry models cost-effectively. GPT 4.1 and GPT 4.1-nano are now available as the latest models from Azure OpenAI, with GPT 4.1 having a 1 million token context limit. The “nano” variant is designed for cost-conscious deployments.

Integration Checklist for .NET Developers

If you’re building an agent on .NET with Azure:

  1. Update AI Toolkit. Install v0.28.1+ from the VS Code Marketplace.
  2. Define your Agent Skills. Use the open Agent Skills standard (compatible with Copilot Chat) to expose your domain logic.
  3. Test locally with Windows ML profiling. Before deploying to Azure, profile memory and latency right in the editor.
  4. Deploy via Agent Framework. Agent Framework is the recommended approach for .NET apps that need to build agentic AI systems with advanced orchestration, multi-agent collaboration, and enterprise-grade security and observability. Sample skill definition (conceptual): 
    public class SalesAgentSkill : IAgentSkill
{
 public string Name => "sales-lookup";
 public async Task<string> ExecuteAsync(string input)
 {
     // Query your CRM, return structured data
     return await _crmClient.SearchLeadsAsync(input);
 }
}

The Bigger Picture: Pragmatism Over Hype

Large language models are great at generalizing knowledge, but many experts say the next wave of enterprise AI adoption will be driven by smaller, more agile language models that can be fine-tuned for domain-specific solutions, with fine-tuned SLMs becoming a staple used by mature AI enterprises in 2026. The AI Toolkit’s shift to Agent Skills and support for GPT 4.1-nano reflects this trend. You’re no longer forced to pay for a 70B-parameter model when a 4B model fine-tuned to your domain will do.

Further Reading

  • https://techcommunity.microsoft.com/blog/azuredevcommunityblog/%F0%9F%9A%80-ai-toolkit-for-vs-code-january-2026-update/4485205
  • https://learn.microsoft.com/en-us/dotnet/ai/dotnet-ai-ecosystem
  • https://learn.microsoft.com/en-us/azure/ai-foundry/openai/whats-new?view=foundry-classic
  • https://devblogs.microsoft.com/dotnet/announcing-dotnet-10/
  • https://learn.microsoft.com/en-us/visualstudio/releases/2026/release-notes
  • https://techcrunch.com/2026/01/02/in-2026-ai-will-move-from-hype-to-pragmatism/