Show HN: Toolnexus for .NET – MCP and agent skills as tools for any LLM

dotnet add package Toolnexus --version 0.1.2
                    

                            
NuGet\Install-Package Toolnexus -Version 0.1.2
                    

                            

Install-Package.

<PackageReference Include="Toolnexus" Version="0.1.2" />
                    

                            

PackageReference, copy this XML node into the project file to reference the package.

<PackageVersion Include="Toolnexus" Version="0.1.2" />
                    

                            Directory.Packages.props
                        

                            
<PackageReference Include="Toolnexus" />
                    

                            Project file
                        

                            

Central Package Management (CPM), copy this XML node into the solution Directory.Packages.props file to version the package.

paket add Toolnexus --version 0.1.2
                    

                            

maintainersfor support.

#r "nuget: Toolnexus, 0.1.2"
                    

                            
#:package Toolnexus@0.1.2
                    

                            
#addin nuget:?package=Toolnexus&version=0.1.2
                    

                            Install as a Cake Addin
                        

                            
#tool nuget:?package=Toolnexus&version=0.1.2
                    

                            Install as a Cake Tool
                        

                            

maintainersfor support.

toolnexus #

Build an agent in a few lines. Point at an mcp.json

and a skills/

folder, call RunAsync()

, and you have a working agent — MCP servers, agent skills, your own functions, and HTTP endpoints unified as one tool set, driving any LLM.

Right-sized.Not a framework (no builders, no config to wade through), not a toy that falls over the moment you need streaming or a retry. Everything a real agent needs — the loop, hooks, streaming, retries, memory — and nothing it doesn't.

The C#/.NET port of toolnexus — the same library, byte-identical, also in JavaScript, Python, Go, and Java. Built on the official ModelContextProtocol

SDK. Targets .NET 10.

Install

dotnet add package Toolnexus

An agent in 3 steps

using Toolnexus;

// 1. tools from an mcp.json + a skills/ folder
await using var tk = await Toolkit.CreateAsync(new Toolkit.Options()
    .WithMcpConfig("./mcp.json")
    .WithSkillsDir("./skills"));

// 2. point at any OpenAI- or Anthropic-style endpoint
var agent = LlmClient.Create(new LlmClient.Options
{
    BaseUrl = "https://openrouter.ai/api/v1",
    Style   = "openai",                 // or "anthropic"
    Model   = "openai/gpt-4o-mini",
    ApiKey  = Environment.GetEnvironmentVariable("OPENROUTER_API_KEY"),
});

// 3. run — skills injected into the system prompt, tools called for you, looped to an answer
var res = await agent.RunAsync("Refund order 1234 for the customer.", tk);
Console.WriteLine(res.Text);

The loop runs call → execute tools → feed back → repeat, with hooks, streaming, retries/backoff, and conversation memory available. res

carries Text

, ToolCalls

, usage, turns, and model.

Add your own tools

using Toolnexus;

// a method → a tool (attribute-based)
public sealed class MathTools
{
    [ToolMethod("add", "Add two numbers")]
    public string Add([Param("a")] double a, [Param("b")] double b) => (a + b).ToString();
}

tk.Register(Tools.FromObject(new MathTools()).ToArray());

// a REST endpoint → a tool
tk.Register(HttpTool.Of(new HttpTool.Options
{
    Name = "create_ticket", Description = "Create a ticket", Method = "POST",
    Url = "https://api.example.com/tickets",
    Headers = new Dictionary<string, string> { ["Authorization"] = "Bearer ${API_TOKEN}" }, // ${ENV} expands, never logged
    InputSchema = new Dictionary<string, object?>
    {
        ["type"] = "object",
        ["properties"] = new Dictionary<string, object?> { ["title"] = new Dictionary<string, object?> { ["type"] = "string" } },
        ["required"] = new List<object?> { "title" },
    },
}));

You can also pass ExtraTools

/ AnnotatedObjects

straight into Toolkit.Options

.

Bring your own loop

var tools  = tk.ToOpenAI();      // or tk.ToAnthropic() / tk.ToGemini()
var system = tk.SkillsPrompt();  // skills catalog for your system prompt
// when the model returns a tool call (name, arguments):
var res = await tk.ExecuteAsync(name, args);   // -> ToolResult(Output, IsError, Metadata)

The four sources

All four appear as one uniform ITool

in tk.Tools()

.

API

More

Full docs, the other four language ports, the shared behavior spec, and runnable examples: https://github.com/muthuishere/toolnexus

MIT licensed.

Product | Versions Compatible and additional computed target framework versions. | |---|---| | .NET | net10.0 is compatible. net10.0-android was computed. net10.0-browser was computed. net10.0-ios was computed. net10.0-maccatalyst was computed. net10.0-macos was computed. net10.0-tvos was computed. net10.0-windows was computed. |

Learn more about

Target Frameworksand.NET Standard.-

net10.0

ModelContextProtocol(>= 1.4.0)

NuGet packages

This package is not used by any NuGet packages.

GitHub repositories

This package is not used by any popular GitHub repositories.