AI Starter Kit for .NET

Your AI agent writes secure, production-ready .NET APIs

Start a new project from an AI-ready template and build real features on day 1.

30 tested AI skills, 3 AI-ready project templates and the AI Code Review Blueprint. Pay once and keep them.

Works with Claude Code, Codex, GitHub Copilot and Cursor · 14-day money-back guarantee

Why skills

AI agents write code that compiles but isn't how your team builds it. A skill teaches the agent once.

Install once. Every agent follows it, every time.

Each skill tested on all four.

One skill folder

.claude/skills/dotnet-tests-review/

Copy once. About 2 minutes.

  • Claude CodeTested
  • CodexTested
  • GitHub CopilotTested
  • CursorTested

What's inside

  • 30 AI skills for .NET

    Your agent writes features, tests, security and production setup the way a senior .NET developer would, in your project's style.

  • 3 AI-ready project templates

    Vertical Slices, Modular monolith and Microservices, with AGENTS.md, CLAUDE.md and every skill already inside.

  • AI Code Review Blueprint

    The checklist a review agent runs on every AI-written pull request: security, data access and quality, with the 5 areas to read yourself.

The 30 skills

Write code the way your team does (5)

Your agent adds features, entities and tests that look like the rest of your codebase.

  • dotnet-add-feature

    Adds an endpoint in the right folder, with your Result and error types, validation and auth, build green.

  • dotnet-add-entity-and-migration

    Adds the entity, its configuration and the migration where they belong, and stops a migration that would delete data.

  • dotnet-result-pattern-errors

    Uses Result<T> for business errors with stable codes and one mapping to ProblemDetails.

  • dotnet-add-permission-policy

    Adds a permission end to end: the constant, the policy, the role that gets it and the endpoint that needs it.

  • dotnet-agents-md-setup

    Writes a short AGENTS.md every agent reads, with commands that were actually run.

Test and review (5)

Tests that catch real bugs, and reviews that catch what AI-written code gets wrong.

  • dotnet-write-integration-tests

    Writes API tests against a real database in Docker that cover every error, 401, 403 and another user's data.

  • dotnet-write-unit-tests

    Writes fast xUnit tests for validators, entity rules and Result error codes.

  • dotnet-tests-review

    Lists the missing tests in priority order, with the file and the assertion for each.

  • dotnet-clean-code-review

    Reviews C# and the build guardrails, and catches warnings switched off to sneak a build through.

  • dotnet-efcore-performance-review

    Finds N+1 queries, ToList before Where, missing projections and AsNoTracking.

Secure your API (5)

The security mistakes AI agents make most, fixed before they ship.

  • dotnet-add-jwt-authentication

    Adds JWT auth with strict token validation and a key from secrets checked at startup.

  • dotnet-resource-authorization

    Stops "I can open someone else's order": every query by id filters by the owner.

  • dotnet-api-security-hardening

    Hardens an API for launch: HSTS, forwarded headers, CORS, request size limits and safe uploads.

  • dotnet-secrets-and-config

    Keeps secrets out of code, images and logs, and stops the app at startup on bad config.

  • dotnet-add-rate-limiting

    Adds rate limits per user, API key or real client IP, even behind a load balancer.

Run in production (4)

Retries, caching and observability set up the way they should be.

  • dotnet-add-resilience

    Adds retries, timeouts and circuit breakers to outgoing calls, without retrying unsafe writes.

  • dotnet-opentelemetry-observability

    Adds OpenTelemetry traces, metrics and logs, linked together, without Aspire.

  • dotnet-select-caching-strategy

    Picks the caching strategy per kind of data and clears the cache after every write.

  • dotnet-select-caching-type

    Picks and wires the right cache: memory, HybridCache with Redis, or output caching.

Aspire (2)

Local orchestration and one dashboard for logs, traces and metrics.

  • dotnet-add-aspire

    Adds Aspire to an existing solution: an AppHost with your databases and queues, wired to the code you have.

  • dotnet-aspire-observability

    Gets logs, traces and metrics into the Aspire dashboard through ServiceDefaults.

Modular monolith (5)

Modules that stay separate as the codebase grows. Works with the Modular monolith template.

  • dotnet-add-module

    Adds a new module the way your existing modules are built, with its own schema and first migration.

  • dotnet-cross-module-communication

    Makes modules talk through public APIs and in-process events, never each other's internals.

  • dotnet-cross-module-queries

    Reads another module's data without touching its tables.

  • dotnet-cross-module-transactions

    Keeps an operation consistent when it writes to several modules.

  • dotnet-module-boundary-tests

    Writes architecture tests that fail the build when modules start reaching into each other.

Ship it (4)

From a working API to a container, a pipeline and background work.

  • dotnet-add-openapi-docs

    Adds OpenAPI docs with Scalar, real examples and auth, so your API documents itself.

  • dotnet-containerize

    Writes a Dockerfile and docker-compose for your API and its database.

  • dotnet-github-actions-ci

    Writes a GitHub Actions pipeline that builds, tests and publishes your app.

  • dotnet-add-background-jobs

    Picks BackgroundService, Hangfire or Quartz for the job, and wires it up.

Tested with and without the skill

The same task, the same agent, scored on the skill's own checklist. Three skills from the kit:

  • dotnet-tests-review

    7/7with the skill, 3/7 without

    Tells you which tests are missing, in priority order, with the file and the assertion for each.

    Without: No priorities, every gap pushed into HTTP tests, and nothing written down.

  • dotnet-agents-md-setup

    7/7with the skill, 4/7 without

    Writes a short AGENTS.md every agent reads: 75 lines, and every command in it was actually run.

    Without: 159 lines, a seeded password copied in, and the integration tests never run.

  • dotnet-clean-code-review

    10/10with the skill, 6.5/10 without

    Reviews AI-written C# and catches the NoWarn, #pragma or disabled analyzer used to sneak a build through.

    Without: Grouped findings by theme, changed the output format nobody asked to change, and missed the tricks.

Templates and Blueprint

3 AI-ready .NET 11 templates

  • Vertical Slices

    One monolith with features organised by business capability. The fastest way to start a new API.

  • Modular monolith

    Loosely coupled modules deployed as one app, talking through public APIs and events.

  • Microservices

    4 services with their own databases behind a YARP gateway, with sync and async communication.

Each one ships with AGENTS.md, CLAUDE.md and all 30 skills, so your agent follows the architecture from the first session.

AI Code Review Blueprint

AI-written code compiles, passes the tests and looks clean, which is what makes it hard to review. The Blueprint is the checklist I give a review agent on every AI-written pull request: security, data access and quality, with .NET examples.

It also tells you the 5 areas to read yourself, line by line: auth, data ownership, migrations, secrets, and JWT and CORS settings.

AI Starter Kit for .NET

One-time payment, no subscription

  • 30 AI skills for .NET
  • 3 AI-ready project templates
  • AI Code Review Blueprint
  • Works with Claude Code, Codex, Copilot and Cursor

14-day money-back guarantee

14-day money-back guarantee. Download everything and try it on your project. If it isn't for you, email info@antondevtips.com within 14 days for a full refund.

Frequently Asked Questions

AI Starter Kit

Still have a question? Email me at info@antondevtips.com.

Anton Martyniuk

Author

Anton Martyniuk — Microsoft MVP, Software Architect, .NET Developer

Hey, I'm Anton — a Microsoft MVP, Software Architect, and .NET developer with over 13 years of experience building production systems, leading teams, and shipping software that actually matters.

I run the Anton DevTips newsletter — one of the top .NET newsletters with 28,000+ subscribers — and share daily content on LinkedIn with a community of 120K+ developers. I've spent years helping .NET developers at every level get unstuck, close knowledge gaps, and reach the next stage of their careers.

I use these skills and templates on my own .NET projects every day. The kit is the part of my AI setup that helps most on day one, in one download.