dotnet-artisan — .NET AI Development Agent

What is dotnet-artisan?

A complete .NET development agent system

It's not scattered tools

It's a complete agent system. A decision-maker analyzes requirements, captures domain vocabulary, designs architecture, then routes to the right skills.

→

Full development lifecycle

From building APIs to debugging crashes: build, test, review, security, debug, CI/CD, upgrade. Install and go.

⟡

Built upon dotnet/skills and dotnet-artisan

Integrates best practices from Microsoft and the community. Maintained by fenzel.

How It Works

7-step orchestration by the decision-maker

Detect .NET Version

Read .csproj / global.json to determine target framework

Align Requirements

Ask clarifying questions, capture domain vocabulary

Load C# Standards

Async, DI, LINQ, serialization best practices

Design Architecture

Invoke architect: choose pattern, structure, build config

Parallel Routing

Decompose into domains, route to matching skills

Specialist Agents

Call domain experts for deep analysis on demand

Learning & Memory

Capture corrections, generalize rules, store in MEMORY.md

11 Skills

6 categories · 173 reference files

Gateway

using-dotnet

Detects .NET intent, triggers decision-maker

—

Gateway

dotnet-advisor

Align → architect → route

2 refs

Baseline

dotnet-csharp

C# standards, async/await, DI, LINQ

27 refs

Build

dotnet-api

Backend API, EF Core, gRPC, SignalR, security

33 refs

Build

dotnet-ui

Blazor, MAUI, WPF, WinUI, Uno

20 refs

Verify

dotnet-testing

xUnit, integration, Playwright, benchmarks

14 refs

Verify

dotnet-debugging

WinDbg / dotnet-dump crash diagnostics

17 refs

Operate

dotnet-devops

CI/CD, containers, migration, Git workflow

19 refs

Operate

dotnet-tooling

Project structure, AOT, CLI, performance, quality, templates

41 refs

Augment

dotnet-ai

MCP servers, Semantic Kernel, RAG

—

Augment

dotnet-workflow

Parallel workflows, context management, verification

—

14 Agents

Role experts + Tool specialists

Role Agents

Work like human experts

Tool Agents

Focus on specific domains

Example Scenarios

The decision-maker in action

1. Building an Order System

You: I need an order management system

AI: What kind of orders?

You: Customer orders. We might split later.

→ 4 rounds → domain-analyst → Event Storming → domain doc → parallel routing

2. Debugging a Crash (OOM)

You: OOM in production, here's the dump file

→ dotnet-debugging → !analyze -v → !dumpheap -stat → !gcroot

3. Learning a Convention

You: Remember, use TimeProvider

→ learning-agent → generalize → deduplicate → write to MEMORY.md

Get Started

Install and go. Zero configuration.

claude plugins marketplace add fenzel999/dotnet-artisan
claude plugins install dotnet-artisan

Compatible with GitHub Copilot, VS Code, and Cursor