semantic-kernel-mcp-server

alkampfergit/semantic-kernel-mcp-server

3.1

If you are the rightful owner of semantic-kernel-mcp-server and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to dayong@mcphub.com.

A sample Model Context Protocol (MCP) server implemented in .NET and C# that demonstrates integration with Microsoft Semantic Kernel agents.

Tools
2
Resources
0
Prompts
0

Semantic Kernel MCP Server

A sample Model Context Protocol (MCP) server implemented in .NET and C# that demonstrates integration with Microsoft Semantic Kernel agents.

This repository provides:

  • An MCP server implemented using the ModelContextProtocol SDK
  • Semantic Kernel integration with a ChatCompletionAgent and a sample TaskManagementPlugin
  • MCP tools exposed as stdio tools that work with VS Code/GitHub Copilot or other MCP clients
  • In-memory task management with CRUD tools, and an optional AI-powered natural language tool

Quick links:

  • Project root: this README
  • Server project: SemanticKernelMcpServer/
  • Server entry point: SemanticKernelMcpServer/Program.cs
  • Tools: SemanticKernelMcpServer/Tools/ (includes SemanticKernelAgentTools.cs and RandomNumberTools.cs)
  • Plugin: SemanticKernelMcpServer/Plugins/TaskManagementPlugin.cs
  • VS Code MCP config: .vscode/mcp.json
  • MCP server metadata: SemanticKernelMcpServer/.mcp/server.json
  • Documentation: QUICK_START.md, IMPLEMENTATION_GUIDE.md, ARCHITECTURE.md, PROJECT_SUMMARY.md

Quick Start

  1. Build the project:
cd SemanticKernelMcpServer
dotnet build
  1. Run locally for testing with VS Code or the MCP client:
dotnet run --project SemanticKernelMcpServer/SemanticKernelMcpServer.csproj
  1. Open VS Code and enable Agent Mode (GitHub Copilot); the server will appear in Select tools.

Enable AI (Optional)

To enable AI-driven features, configure an AI provider in Tools/SemanticKernelAgentTools.cs (Azure OpenAI or OpenAI) and set the following environment variables:

  • AZURE_OPENAI_ENDPOINT and AZURE_OPENAI_DEPLOYMENT and AZURE_OPENAI_API_KEY, or
  • OPENAI_API_KEY for OpenAI

See QUICK_START.md for detailed instructions and examples.

Contributing

Contributions are welcome—open issues and PRs for bugs, improvements, or ideas.

License

MIT — see for details.

If you'd like, I can now:

  • Add a license file and GitHub Actions CI
  • Rename master to main and update default branch
  • Create a release and package to NuGet (MCP server) for distribution

Tell me which you'd like next and I'll continue.