Generate Testability Wrappers logo

Generate Testability Wrappers

OrganizationPopular
dotnet
generate-testability-wrappers

DO NOT USE when the target already consumes an injected interface or built-in abstraction such as IFileSystem or TimeProvider, even if the request says "generate a wrapper"; no new wrapper is needed. Use only when C# source calls an ambient/static dependency and no injectable seam exists: first-time TimeProvider, IHttpClientFactory, or System.IO.Abstractions adoption; minimal Environment/Console/Process wrappers; IProcessRunner; DI registration; or an ambient seam that preserves a static API. Exclude static detection (detect-static-dependencies), migration to an existing/registered abstraction (migrate-static-to-wrapper), one blocked behavior plus deterministic tests (testability-obstacle), and general interface design.

Overview

Publisherdotnet
Repositoryskills
Skill namegenerate-testability-wrappers
Stars
5.5K
Forks
420
Bundled files
Instructions only
LicenseMIT
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 dotnet on GitHub. Read the source before you install it.

Installation

Install the Generate Testability Wrappers 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.
    https://github.com/dotnet/skills/tree/main/plugins/dotnet-test/skills/generate-testability-wrappers
  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/dotnet/skills.git /tmp/skills
mkdir -p .claude/skills
cp -r /tmp/skills/plugins/dotnet-test/skills/generate-testability-wrappers .claude/skills/generate-testability-wrappers
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Generate Testability Wrappers 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 Generate Testability Wrappers 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 Generate Testability Wrappers 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.

Generate Testability Wrappers

Generate wrapper interfaces, default implementations, and DI service registration code for untestable static dependencies. For statics that already have .NET built-in abstractions (TimeProvider, IHttpClientFactory), guide adoption of the built-in. For statics without built-in alternatives, generate custom minimal wrappers.

When to Use

  • After running detect-static-dependencies and identifying which statics to wrap
  • When the user asks to make a class testable by replacing statics with injected abstractions
  • When adopting TimeProvider (.NET 8+) or System.IO.Abstractions
  • When creating a custom wrapper for Environment.*, Console.*, or Process.*
  • When a released static API needs an ambient seam because signatures cannot change

When Not to Use

  • The user wants to find statics first (use detect-static-dependencies)
  • The user wants to bulk-replace call sites (use migrate-static-to-wrapper)
  • The static is already behind an interface

If the target already consumes an injected interface or built-in abstraction, stop: do not add a second wrapper, project, registration, or test around that seam.

A missing DI package does not by itself force an ambient seam. For an instantiable class, prefer constructor injection and compose it explicitly or show the requested registration. Use Step 5 when the API is static and its signatures must stay static, or when the user explicitly forbids caller construction/DI changes.

Inputs

InputRequiredDescription
Static categoryYesWhich category: time, filesystem, environment, network, console, process
Target frameworkYesThe TargetFramework from .csproj (affects which built-in abstractions exist)
CompositionNoExisting DI framework, explicit/manual construction, or immutable static API
NamespaceNoTarget namespace for generated wrapper code

Workflow

Step 1: Determine the abstraction strategy

Based on the category and target framework:

Category.NET 8+.NET 6-7.NET Framework
TimeTimeProvider (built-in)TimeProvider via Microsoft.Bcl.TimeProvider NuGetCustom ISystemClock
File systemSystem.IO.Abstractions (NuGet)SameSame
HTTPIHttpClientFactory (built-in)SameSame
EnvironmentCustom IEnvironmentProviderSameSame
ConsoleCustom IConsoleSameSame
ProcessCustom IProcessRunnerSameSame

The table picks which abstraction. How it reaches the code under test is a separate axis:

  • instantiable class: constructor injection, even if current callers compose the object manually;
  • existing container: add compile-ready registration following its conventions;
  • public static API/signatures that cannot change: Step 5's ambient seam.

Check for a host builder, IServiceCollection, existing registrations, and construction sites. Do not infer "must remain static" merely because the project currently has no container.

Step 2: Generate built-in abstraction adoption (Time, HTTP)

TimeProvider (.NET 8+)

No wrapper code needed. Complete all four parts: production registration, constructor injection, a FakeTimeProvider test, and the testing package.

  1. Register in DI:
csharp
builder.Services.AddSingleton(TimeProvider.System);
  1. Inject into classes:
csharp
public class OrderProcessor(TimeProvider timeProvider)
{
    public bool IsExpired(Order order)
        => timeProvider.GetUtcNow() > order.ExpiresAt;
}
  1. Test with FakeTimeProvider:
csharp
// Requires Microsoft.Extensions.TimeProvider.Testing NuGet
var fakeTime = new FakeTimeProvider(new DateTimeOffset(2026, 1, 15, 0, 0, 0, TimeSpan.Zero));
var processor = new OrderProcessor(fakeTime);
fakeTime.Advance(TimeSpan.FromDays(1));
Assert.True(processor.IsExpired(order));

The assertion must prove a time-dependent result after the fake is pinned or advanced. Merely constructing FakeTimeProvider is not a test. When the project has no container but the target is an instantiable class, inject TimeProvider anyway and show explicit production construction with TimeProvider.System; do not replace it with a custom static clock.

Before calling the adoption complete, verify the repository contains or the answer supplies every required artifact: the testing package reference, the production composition/registration, every affected constructor call, and a runnable fake-time test. A code fragment that omits one of those integration points is guidance, not a completed adoption.

TimeProvider (pre-.NET 8)

Guide: install Microsoft.Bcl.TimeProvider NuGet. Same API as above.

IHttpClientFactory

No wrapper code needed. Register a typed client via builder.Services.AddHttpClient<MyService>() and inject HttpClient directly into the class constructor. Preserve cancellation by passing the caller's token to the HTTP operation.

For tests, provide a complete fake HttpMessageHandler whose SendAsync returns a deterministic HttpResponseMessage, construct HttpClient with that handler, and exercise the typed client without network access. Compile and run the focused test when the task asks for implementation; do not stop at a schematic handler method.

When production uses a typed-client registration, test that same registration pipeline: configure its primary handler in ServiceCollection, resolve the typed client, and call it. A test that manually constructs HttpClient proves the class but not the DI registration the task asked to adopt.

Step 3: Generate custom wrappers (Environment, Console, Process)

For categories without built-in abstractions, follow this template:

Interface — define the minimal surface

Only include methods that were actually detected in the codebase. Do NOT generate a wrapper for every possible member — wrap only what is used.

Prefer a stateless operation-shaped interface. For example, if a caller only needs to start a process, wait, and return its exit code, expose one Run operation rather than a stateful wrapper that leaks Process lifecycle.

csharp
namespace <Namespace>;

/// <summary>
/// Abstraction over <static class> for testability. 
/// </summary>
public interface I<WrapperName>
{
    // One method per detected static call
    <return type> <MethodName>(<parameters>);
}
Default implementation — delegate to the real static
csharp
namespace <Namespace>;

/// <summary>
/// Default implementation that delegates to <static class>.
/// </summary>
public sealed class <WrapperName> : I<WrapperName>
{
    public <return type> <MethodName>(<parameters>)
        => <StaticClass>.<Method>(<arguments>);
}
DI registration
csharp
// In Program.cs or Startup.cs:
builder.Services.AddSingleton<I<WrapperName>, <WrapperName>>();

Treat registration as a deliverable, not a sentence in the summary. Add it to the repository's existing registration surface when one exists. Otherwise show the exact compile-ready statement and identify where the caller should place it. Stateless delegating wrappers are singleton; if state must be retained, explain why a shorter lifetime is required.

Step 4: Generate file system wrapper adoption

Prefer the established System.IO.Abstractions NuGet package over custom wrappers:

  1. Install the package:
dotnet add package System.IO.Abstractions
  1. Register in DI:
csharp
builder.Services.AddSingleton<IFileSystem, FileSystem>();
  1. Inject IFileSystem into classes:
csharp
public class ConfigLoader(IFileSystem fileSystem)
{
    public string LoadConfig(string path)
        => fileSystem.File.ReadAllText(path);
}
  1. Test with MockFileSystem:
dotnet add <TestProject> package System.IO.Abstractions.TestingHelpers
csharp
var mockFs = new MockFileSystem(new Dictionary<string, MockFileData>
{
    { "/config.json", new MockFileData("{\"key\": \"value\"}") }
});
var loader = new ConfigLoader(mockFs);
Assert.Equal("{\"key\": \"value\"}", loader.LoadConfig("/config.json"));

Package-first adoption is exclusive: add both package references, use IFileSystem in production, register or explicitly compose FileSystem, and seed MockFileSystem before exercising the consumer. Do not also generate a second custom filesystem interface, and do not present an unseeded mock whose test could pass without proving the requested read/write behavior.

Step 5: Generate a signature-preserving ambient context

Use this pattern when the API must remain static or its released signatures cannot accept a dependency:

csharp
public static class Clock
{
    private static readonly AsyncLocal<Func<DateTime>?> s_override = new();
    public static DateTime UtcNow
        => s_override.Value?.Invoke() ?? TimeProvider.System.GetUtcNow().UtcDateTime;

    internal static IDisposable Override(DateTime fixedUtcTime)
    {
        if (fixedUtcTime.Kind != DateTimeKind.Utc)
            throw new ArgumentException("The override must be UTC.", nameof(fixedUtcTime));

        var previous = s_override.Value;
        s_override.Value = () => fixedUtcTime;
        return new Scope(previous);
    }
    private sealed class Scope : IDisposable
    {
        private readonly Func<DateTime>? _previous;
        private bool _disposed;

        public Scope(Func<DateTime>? previous)
        {
            _previous = previous;
        }

        public void Dispose()
        {
            if (_disposed)
                return;

            s_override.Value = _previous;
            _disposed = true;
        }
    }
}

Key trade-offs: AsyncLocal<T> ensures parallel tests don't interfere; production cost is one null check per call; the static readonly field is essentially free.

Three properties this pattern must keep, because each has broken a real migration:

  • Scope the override and make it reversible. Return an IDisposable that restores the previous value, so a test cannot leak a pinned time into the next one. A bare setter, or a manual try/finally at each call site, puts that burden on every test author.
  • Use AsyncLocal<T>, never [ThreadStatic]. [ThreadStatic] does not flow across await, so the override silently disappears mid-test.
  • Preserve the semantics of the member you are replacing. Substituting DateTime.UtcNow with a local-time source changes the DateTimeKind every existing caller and stored value depends on — pair UtcNow with GetUtcNow(), and Now with GetLocalNow().
  • Prove the test assembly can reach the override. An internal override is inaccessible from a separate test assembly unless the production project adds the exact InternalsVisibleTo for that test assembly. Otherwise use an already-public seam only when a public API change is authorized. Never show a test calling an inaccessible member.

The same shape works for non-time statics: swap TimeProvider.System.GetUtcNow() for the real static call and keep the override slot, the disposable scope, and the original semantics.

Step 6: Place generated files

Generate files following the project's existing conventions:

  • If there is an Abstractions/ or Interfaces/ folder, place the interface there
  • If there is an Infrastructure/ or Services/ folder, place the implementation there
  • Otherwise, create files next to the code that uses the static

Always generate:

  1. The interface file (or adoption instructions for built-in abstractions)
  2. The default implementation file
  3. The compile-ready DI registration, applied to an existing registration surface or shown at the exact composition point.
  4. A deterministic substitution example or focused test that exercises the consumer without the ambient resource.

Skip registration entirely on the ambient-seam path: there is no container to register into, and offering one anyway is the failure mode that made a user ask for the seam in the first place.

Before reporting completion, verify the delivered output contains every item the prompt requested. In particular, do not summarize "singleton registration" when no registration code was added or shown, and do not claim testability without demonstrating how the consumer receives a fake. For console wrappers, exercise the consumer with a fake that both captures the prompt and supplies the returned input; a build or banner-only run does not prove the prompt flow.

Validation

  • Generated interface only wraps statics that were actually detected (not speculative)
  • Default implementation delegates to the real static with no behavior changes
  • DI registration uses AddSingleton for stateless wrappers, AddTransient for stateful ones
  • NuGet packages are recommended where established libraries exist (System.IO.Abstractions, etc.)
  • For .NET 8+, TimeProvider is recommended over custom ISystemClock
  • TimeProvider adoption includes injection, production composition, FakeTimeProvider, its testing package, and an assertion on a time-dependent result
  • HTTP adoption includes a complete fake-handler test and preserves cancellation
  • On injection paths, registration or explicit composition is compile-ready, and a fake demonstrates the consumer without the real ambient dependency
  • Ambient context pattern includes AsyncLocal<T>, a scoped IDisposable that restores the previous value, and trade-off explanation
  • On the ambient-seam path, no IServiceCollection registration is proposed, the separate test assembly can reach the override, and the replaced member's return type and semantics (UtcNow vs Now, and its DateTimeKind) are preserved

Common Pitfalls

PitfallSolution
Treating "no DI package" as "must be ambient"Inject into instantiable classes and compose explicitly; reserve Step 5 for static/signature-preserving APIs
Wrapping ALL members of a static classOnly wrap methods actually called in the codebase
Custom time wrapper on .NET 8+Use built-in TimeProvider instead
Custom file system wrapperPrefer System.IO.Abstractions NuGet — battle-tested, complete
Registering scoped when singleton sufficesStateless wrappers should be AddSingleton
Forgetting test helper packagesMicrosoft.Extensions.TimeProvider.Testing for time, System.IO.Abstractions.TestingHelpers for filesystem
Ambient context without AsyncLocalNon-async [ThreadStatic] breaks with async/await — always use AsyncLocal<T>
Showing an internal ambient override to external testsAdd the exact friend assembly or use an authorized public seam; compile the test project

Frequently asked questions

What does the Generate Testability Wrappers AI skill do?

DO NOT USE when the target already consumes an injected interface or built-in abstraction such as IFileSystem or TimeProvider, even if the request says "generate a wrapper"; no new wrapper is needed. Use only when C# source calls an ambient/static dependency and no injectable seam exists: first-time TimeProvider, IHttpClientFactory, or System.IO.Abstractions adoption; minimal Environment/Console/Process wrappers; IProcessRunner; DI registration; or an ambient seam that preserves a static API. Exclude static detection (detect-static-dependencies), migration to an existing/registered abstract...

Why use Generate Testability Wrappers on TypingMind?

Because you install it once and use it with any model. Generate Testability Wrappers 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 Generate Testability Wrappers in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/dotnet/skills/tree/main/plugins/dotnet-test/skills/generate-testability-wrappers. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Generate Testability Wrappers?

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 Generate Testability Wrappers?

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

Is the Generate Testability Wrappers AI skill free?

Yes. It is published on GitHub by dotnet under the MIT license. 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 👇