iOS architecture skills for Claude Code and AI coding agents

Coding agents like Claude Code can write a SwiftUI feature fast. The problem is what the codebase looks like after ten of them. These skills tell the agent how to split a feature, where storage goes (GRDB), how to test async code, and how to check the screen at the end.

The toolkit covers SwiftUI, Clean Architecture, MVVM, GRDB, and Swift concurrency. Its skills divide responsibilities across architecture, reusable UI, platform integration, storage, testing, and visual checks.

What the toolkit covers

Use the architecture skill for a new app or feature. Use the persistence skill for a database change, and the testing skill for concurrency and dependency injection. Pick the one that matches the size of the change.

Start with one feature

Choose a small feature and describe its expected behavior, storage needs, and existing project conventions. Follow the repository's installation instructions, then inspect the generated changes and verify the result in your project.

Questions

Is this an application framework? It is a toolkit of instructions and supporting resources for coding agents.

Does it include persistence guidance? Yes. The public toolkit includes focused GRDB workflows alongside broader feature architecture.

Does it replace code review? Review and verification remain part of using generated changes in a real project.

Read the repository and installation guide.

Try it on a bounded feature

Install the toolkit using the repository instructions, then start with a request that states behavior and constraints:

Add a saved-items feature to this SwiftUI app. Persist items locally, follow the existing feature boundaries, and verify that saved items survive relaunch.

Review the domain model, storage boundary, and view state separately. The result still needs to compile and be exercised in your app.

For the visual check, I use screenshots-ios. All projects →