Dotnet Csharp Source Generators logo

Dotnet Csharp Source Generators

Community
wshaddix
dotnet-csharp-source-generators

Creating source generators. IIncrementalGenerator, GeneratedRegex, LoggerMessage, STJ source-gen.

Overview

Publisherwshaddix
Repositorydotnet-skills
Skill namedotnet-csharp-source-generators
Stars
79
Forks
13
Bundled files
Instructions only
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by wshaddix on GitHub. Read the source before you install it.

Installation

Install the Dotnet Csharp Source Generators AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/wshaddix/dotnet-skills.git /tmp/dotnet-skills
mkdir -p .claude/skills
cp -r /tmp/dotnet-skills/skills/dotnet-csharp-source-generators .claude/skills/dotnet-csharp-source-generators
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Dotnet Csharp Source Generators in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Dotnet Csharp Source Generators on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Dotnet Csharp Source Generators is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

dotnet-csharp-source-generators

Guidance for both creating and consuming Roslyn source generators in .NET. Creating: IIncrementalGenerator, syntax providers, semantic analysis, emit patterns, diagnostic reporting, testing with CSharpGeneratorDriver. Consuming: [GeneratedRegex], [LoggerMessage], System.Text.Json source generation, [JsonSerializable].

Cross-references: [skill:dotnet-csharp-modern-patterns] for partial properties and related C# features, [skill:dotnet-csharp-coding-standards] for naming conventions.


Creating Source Generators

Project Setup

Source generators are shipped as analyzers targeting netstandard2.0.

xml
<Project Sdk="Microsoft.NET.Sdk">
  <PropertyGroup>
    <TargetFramework>netstandard2.0</TargetFramework>
    <EnforceExtendedAnalyzerRules>true</EnforceExtendedAnalyzerRules>
    <IsRoslynComponent>true</IsRoslynComponent>
    <LangVersion>latest</LangVersion>
  </PropertyGroup>

  <ItemGroup>
    <PackageReference Include="Microsoft.CodeAnalysis.Analyzers" Version="3.3.4" PrivateAssets="all" />
    <PackageReference Include="Microsoft.CodeAnalysis.CSharp" Version="4.12.0" PrivateAssets="all" />
  </ItemGroup>
</Project>

Always target netstandard2.0. Generators load into the compiler process, which requires this TFM for compatibility. Use LangVersion>latest to write modern C# in the generator itself.

IIncrementalGenerator (Preferred)

Always use IIncrementalGenerator over the legacy ISourceGenerator. Incremental generators are cache-aware and only re-run when inputs change, making them significantly faster in IDE scenarios.

csharp
[Generator]
public sealed class AutoNotifyGenerator : IIncrementalGenerator
{
    public void Initialize(IncrementalGeneratorInitializationContext context)
    {
        // Step 1: Filter syntax nodes to candidate fields
        var fieldDeclarations = context.SyntaxProvider
            .ForAttributeWithMetadataName(
                "MyLib.AutoNotifyAttribute",
                predicate: static (node, _) => node is FieldDeclarationSyntax,
                transform: static (ctx, _) => GetFieldInfo(ctx))
            .Where(static info => info is not null)
            .Select(static (info, _) => info!.Value);

        // Step 2: Group fields by containing type, then emit one file per type
        context.RegisterSourceOutput(fieldDeclarations.Collect(),
            static (spc, fields) => Execute(fields, spc));
    }

    private static FieldInfo? GetFieldInfo(
        GeneratorAttributeSyntaxContext context)
    {
        var fieldSymbol = context.TargetSymbol as IFieldSymbol;
        if (fieldSymbol is null)
            return null;

        var containingType = fieldSymbol.ContainingType;

        // Use fully qualified type name to handle generic and nested types
        var fullTypeName = containingType.ToDisplayString(
            SymbolDisplayFormat.FullyQualifiedFormat
                .WithGlobalNamespaceStyle(SymbolDisplayGlobalNamespaceStyle.Omitted));

        return new FieldInfo(
            fieldSymbol.ContainingNamespace.IsGlobalNamespace
                ? ""
                : fieldSymbol.ContainingNamespace.ToDisplayString(),
            containingType.Name,
            fullTypeName,
            fieldSymbol.Name,
            fieldSymbol.Type.ToDisplayString());
    }

    private static void Execute(
        ImmutableArray<FieldInfo> fields,
        SourceProductionContext context)
    {
        // Group by fully qualified type name to emit one file per class
        foreach (var group in fields.GroupBy(f => f.FullTypeName))
        {
            var first = group.First();
            var ns = first.Namespace;
            var className = first.ClassName;
            var properties = new StringBuilder();

            foreach (var field in group)
            {
                var propertyName = GetPropertyName(field.FieldName);
                properties.AppendLine($$"""
                        public {{field.FieldType}} {{propertyName}}
                        {
                            get => {{field.FieldName}};
                            set
                            {
                                if (!global::System.Collections.Generic.EqualityComparer<{{field.FieldType}}>.Default.Equals({{field.FieldName}}, value))
                                {
                                    {{field.FieldName}} = value;
                                    PropertyChanged?.Invoke(this,
                                        new global::System.ComponentModel.PropertyChangedEventArgs(nameof({{propertyName}})));
                                }
                            }
                        }
                    """);
            }

            // Handle global namespace (no namespace declaration)
            var nsBlock = string.IsNullOrEmpty(ns) ? "" : $"namespace {ns};\n\n";

            var source = $$"""
                // <auto-generated/>
                #nullable enable

                {{nsBlock}}partial class {{className}}
                    : global::System.ComponentModel.INotifyPropertyChanged
                {
                    public event global::System.ComponentModel.PropertyChangedEventHandler? PropertyChanged;

                {{properties}}
                }
                """;

            // Include namespace in hint name to avoid collisions across namespaces
            var hintPrefix = string.IsNullOrEmpty(ns) ? className : $"{ns}.{className}";
            context.AddSource($"{hintPrefix}.AutoNotify.g.cs", source);
        }
    }

    private static string GetPropertyName(string fieldName)
        => fieldName.TrimStart('_') is [var first, .. var rest]
            ? $"{char.ToUpperInvariant(first)}{rest}"
            : fieldName;
}

internal readonly record struct FieldInfo(
    string Namespace,
    string ClassName,
    string FullTypeName,
    string FieldName,
    string FieldType);

Scope note: This example targets top-level, non-generic classes for clarity. A production generator should also handle generic type parameters (emitting matching partial class Foo<T> declarations) and nested types (emitting nested partial class hierarchies). Report a diagnostic for unsupported shapes rather than emitting invalid code.

Key Pipeline Design Rules

  1. Filter early -- Use ForAttributeWithMetadataName or CreateSyntaxProvider with a tight predicate to minimize work.
  2. Transform to simple data -- Extract only the data you need (strings, records) in the transform step. Never pass ISymbol or SyntaxNode through the pipeline (they hold the compilation alive and break caching).
  3. Use value equality -- Pipeline outputs are compared by value. Use record struct or implement IEquatable<T> for custom types.
  4. Emit deterministic output -- Same inputs must produce identical source. Use // <auto-generated/> and #nullable enable headers.

Syntax Providers

csharp
// ForAttributeWithMetadataName -- most common, filters by attribute
var candidates = context.SyntaxProvider.ForAttributeWithMetadataName(
    "MyLib.GenerateMapperAttribute",
    predicate: static (node, _) => node is ClassDeclarationSyntax,
    transform: static (ctx, _) => /* extract info */);

// CreateSyntaxProvider -- general-purpose, any syntax predicate
var candidates = context.SyntaxProvider.CreateSyntaxProvider(
    predicate: static (node, _) => node is MethodDeclarationSyntax m
        && m.Modifiers.Any(SyntaxKind.PartialKeyword),
    transform: static (ctx, _) => /* extract info */);

Diagnostic Reporting

Report errors and warnings through SourceProductionContext rather than throwing exceptions. To report location-specific diagnostics, include a Location in your pipeline data (captured from the syntax node in the transform step).

csharp
private static readonly DiagnosticDescriptor InvalidFieldType = new(
    id: "AN001",
    title: "Invalid field type for AutoNotify",
    messageFormat: "Field '{0}' must be a non-pointer type",
    category: "AutoNotify",
    defaultSeverity: DiagnosticSeverity.Error,
    isEnabledByDefault: true);

// In the transform step, capture location:
var location = context.TargetNode.GetLocation();

// In the Execute method, report with location:
context.ReportDiagnostic(Diagnostic.Create(
    InvalidFieldType,
    location,       // captured from syntax node, not from projected data
    fieldName));

Note: Location is not value-equatable, so including it in your pipeline record breaks incremental caching. A common pattern is to carry it as a separate field that you exclude from equality, or report diagnostics in a CreateSyntaxProvider step before projecting to value types.

Emit Patterns

csharp
// Prefer raw string literals for templates (C# 11+, in the generator project)
var source = $$"""
    // <auto-generated/>
    #nullable enable

    namespace {{ns}};

    partial class {{className}}
    {
        {{generatedMembers}}
    }
    """;

context.AddSource($"{className}.g.cs", source);

File naming convention: {TypeName}.{Feature}.g.cs -- the .g.cs suffix signals generated code and is excluded by many linters.

Post-Init Output (Static Source)

Use RegisterPostInitializationOutput for marker attributes and helper types that do not depend on user code:

csharp
context.RegisterPostInitializationOutput(static ctx =>
{
    ctx.AddSource("AutoNotifyAttribute.g.cs", """
        // <auto-generated/>
        namespace MyLib;

        [System.AttributeUsage(System.AttributeTargets.Field)]
        internal sealed class AutoNotifyAttribute : System.Attribute { }
        """);
});

Testing Source Generators

Use CSharpGeneratorDriver to run generators in-memory and verify output.

csharp
using Microsoft.CodeAnalysis;
using Microsoft.CodeAnalysis.CSharp;

[Fact]
public void Generator_ProducesExpectedOutput()
{
    // Arrange
    var source = """
        using MyLib;

        namespace TestApp;

        public partial class ViewModel
        {
            [AutoNotify]
            private string _name = "";
        }
        """;

    var syntaxTree = CSharpSyntaxTree.ParseText(source);
    var references = AppDomain.CurrentDomain.GetAssemblies()
        .Where(a => !a.IsDynamic && !string.IsNullOrEmpty(a.Location))
        .Select(a => MetadataReference.CreateFromFile(a.Location))
        .Cast<MetadataReference>()
        .ToList();

    var compilation = CSharpCompilation.Create("TestAssembly",
        [syntaxTree],
        references,
        new CSharpCompilationOptions(OutputKind.DynamicallyLinkedLibrary));

    var generator = new AutoNotifyGenerator();

    // Act
    GeneratorDriver driver = CSharpGeneratorDriver.Create(generator);
    driver = driver.RunGeneratorsAndUpdateCompilation(
        compilation, out var outputCompilation, out var diagnostics);

    // Assert
    Assert.Empty(diagnostics.Where(d => d.Severity == DiagnosticSeverity.Error));

    var runResult = driver.GetRunResult();
    Assert.Single(runResult.GeneratedTrees);

    var generatedSource = runResult.GeneratedTrees[0].GetText().ToString();
    Assert.Contains("public string Name", generatedSource);
}

Snapshot Testing (Verify)

For more robust testing, use the Verify.SourceGenerators package to snapshot-test generated output:

csharp
[Fact]
public Task Generator_SnapshotTest()
{
    var source = """
        using MyLib;
        namespace TestApp;
        public partial class ViewModel
        {
            [AutoNotify]
            private string _name = "";
        }
        """;

    return TestHelper.Verify(source);
}

Consuming Built-In Source Generators

[GeneratedRegex] (net7.0+)

Compile-time regex generation. Zero runtime compilation cost, AOT-compatible.

csharp
public partial class Validators
{
    [GeneratedRegex(@"^[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}$",
        RegexOptions.Compiled | RegexOptions.IgnoreCase)]
    private static partial Regex EmailRegex();

    public static bool IsValidEmail(string email)
        => EmailRegex().IsMatch(email);
}

Key rules:

  • Method must be static partial returning Regex
  • Place on partial class (or partial struct)
  • Replaces new Regex(...) with zero allocation at runtime
  • Supports all RegexOptions except RegexOptions.Compiled (which is ignored -- the source generator replaces it)

[LoggerMessage] (net6.0+)

High-performance structured logging with zero-allocation at log-disabled levels.

csharp
public static partial class LogMessages
{
    [LoggerMessage(Level = LogLevel.Information,
        Message = "Processing order {OrderId} for customer {CustomerId}")]
    public static partial void OrderProcessing(
        this ILogger logger, int orderId, string customerId);

    [LoggerMessage(Level = LogLevel.Error,
        Message = "Failed to process order {OrderId}")]
    public static partial void OrderProcessingFailed(
        this ILogger logger, int orderId, Exception exception);
}

// Usage
logger.OrderProcessing(order.Id, order.CustomerId);

Key rules:

  • Methods must be static partial in a partial class
  • Parameters matching {Placeholder} in the message are logged as structured data
  • Exception parameter is logged automatically (do not include in message template)
  • Event IDs are auto-assigned if not specified; specify explicit IDs for stable telemetry

System.Text.Json Source Generation (net6.0+)

AOT-compatible JSON serialization. Eliminates runtime reflection.

csharp
[JsonSerializable(typeof(Order))]
[JsonSerializable(typeof(List<Order>))]
[JsonSerializable(typeof(Customer))]
[JsonSourceGenerationOptions(
    PropertyNamingPolicy = JsonKnownNamingPolicy.CamelCase,
    DefaultIgnoreCondition = JsonIgnoreCondition.WhenWritingNull)]
public partial class AppJsonContext : JsonSerializerContext;
Registration in ASP.NET Core
csharp
builder.Services.ConfigureHttpJsonOptions(options =>
{
    options.SerializerOptions.TypeInfoResolverChain.Insert(0, AppJsonContext.Default);
});

// Or for Minimal APIs
app.MapGet("/orders/{id}", async (int id, IOrderService service) =>
{
    var order = await service.GetByIdAsync(id);
    return order is not null
        ? Results.Ok(order)
        : Results.NotFound();
});
Manual Serialization
csharp
// Serialize
var json = JsonSerializer.Serialize(order, AppJsonContext.Default.Order);

// Deserialize
var order = JsonSerializer.Deserialize(json, AppJsonContext.Default.Order);

// With stream
await JsonSerializer.SerializeAsync(stream, orders,
    AppJsonContext.Default.ListOrder);

Key rules:

  • Register all types that need serialization in [JsonSerializable] attributes
  • Use TypeInfoResolverChain (net8.0+) to combine multiple contexts
  • Required for Native AOT -- reflection-based serialization is trimmed
  • See [skill:dotnet-csharp-modern-patterns] for related C# features used in generated code

[JsonSerializable] with Polymorphism (net7.0+)

csharp
[JsonDerivedType(typeof(CreditCardPayment), "credit")]
[JsonDerivedType(typeof(BankTransferPayment), "bank")]
public abstract class Payment
{
    public decimal Amount { get; init; }
}

public class CreditCardPayment : Payment
{
    public required string CardLast4 { get; init; }
}

public class BankTransferPayment : Payment
{
    public required string AccountNumber { get; init; }
}

[JsonSerializable(typeof(Payment))]
public partial class PaymentJsonContext : JsonSerializerContext;

Generator Reference: Packaging and Consumption

Referencing a Generator in a Consuming Project

xml
<ItemGroup>
  <ProjectReference Include="..\MyGenerator\MyGenerator.csproj"
                    OutputItemType="Analyzer"
                    ReferenceOutputAssembly="false" />
</ItemGroup>

NuGet Package Layout

When shipping a generator as a NuGet package, place the assembly under analyzers/dotnet/cs/:

MyGenerator.nupkg
  analyzers/
    dotnet/
      cs/
        MyGenerator.dll
  lib/
    netstandard2.0/
      _._   (empty placeholder if no runtime dependency)
xml
<!-- In the generator .csproj -->
<PropertyGroup>
  <IncludeBuildOutput>false</IncludeBuildOutput>
  <DevelopmentDependency>true</DevelopmentDependency>
</PropertyGroup>

<ItemGroup>
  <None Include="$(OutputPath)\$(AssemblyName).dll"
        Pack="true"
        PackagePath="analyzers/dotnet/cs" />
</ItemGroup>

Debugging Source Generators

csharp
// Add to Initialize() for attach-debugger workflow
#if DEBUG
if (!System.Diagnostics.Debugger.IsAttached)
{
    System.Diagnostics.Debugger.Launch();
}
#endif

Alternatively, emit generated files to disk for inspection:

xml
<!-- In the consuming project -->
<PropertyGroup>
  <EmitCompilerGeneratedFiles>true</EmitCompilerGeneratedFiles>
  <CompilerGeneratedFilesOutputPath>Generated</CompilerGeneratedFilesOutputPath>
</PropertyGroup>

Add Generated/ to .gitignore.


References

Frequently asked questions

What does the Dotnet Csharp Source Generators AI skill do?

Creating source generators. IIncrementalGenerator, GeneratedRegex, LoggerMessage, STJ source-gen.

Why use Dotnet Csharp Source Generators on TypingMind?

Because you install it once and use it with any model. Dotnet Csharp Source Generators is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Dotnet Csharp Source Generators in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/wshaddix/dotnet-skills/tree/master/skills/dotnet-csharp-source-generators. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Dotnet Csharp Source Generators?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Dotnet Csharp Source Generators?

As many as you like. As long as a model supports skills, you can use Dotnet Csharp Source Generators with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Dotnet Csharp Source Generators AI skill free?

It is published on GitHub by wshaddix. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇