dotnet offers insights into .NET development, C# programming language, and cross-platform application development, providing documentation, tutorials, and best practices for building modern and scalable applications with .NET framework. By exploring dotnet's curated content, developers can learn about .NET Core, ASP.NET, and Entity Framework for developing web APIs, microservices, and desktop applications. Whether you're targeting Windows, Linux, or macOS platforms, dotnet offers resources to accelerate your .NET development workflow and deliver robust and performant applications.

.NET Blog

Microsoft's .NET team has launched dotnet/skills, a GitHub repository hosting agent skills for .NET developers. Agent skills are lightweight packages bundling intent, task-specific context, and supporting artifacts that coding agents like GitHub Copilot and Claude Code can discover and use to complete tasks more reliably. The repo follows the Agent Skills specification and is organized as a plugin marketplace. Skills can be installed via CLI commands or VS Code Copilot settings. The team validates each skill against a baseline to measure whether it actually improves agent responses, and plans to iterate based on community feedback.

Extend your coding agent with .NET Skills

What we are doing with dotnet/skills Copy link

Discovery, installation, and using skills Copy link

Interesting approach. Having task-specific skills for coding agents should make Copilot and similar tools much more reliable, especially if each skill is validated against real tasks before being added.

This is a smart move by the .NET team. The problem with coding agents today isn’t intelligence — it’s context. Models hallucinate or underperform on framework-specific tasks not because they can’t reason, but because they lack precise, task-scoped knowledge at inference time. Bundling intent + context + artifacts into installable skills is essentially retrieval-augmented generation with curation baked in.
What I find most promising is the baseline validation step. Most agent tooling ships on vibes — measuring whether a skill actually improves agent responses before publishing is a higher bar than most plugin ecosystems hold. The real test will be community-contributed skills: can they maintain that validation discipline at scale, or does quality drift as the marketplace grows?

Looks great, although we should need a way of authenticating skills in future