A Swift Concurrency agent skill designed for AI coding assistants like Claude Code, Codex, and Gemini aims to improve the quality of generated Swift concurrency code. It targets common mistakes LLMs make with modern Swift features such as async/await, actors, and task groups, providing a focused set of checks derived from expert knowledge to enhance code safety and efficiency.

How It Works

This project provides an "agent skill" that integrates directly into AI coding tools. It leverages rules and checks developed from extensive research and a dedicated book on Swift concurrency. The skill's design prioritizes addressing specific, nuanced errors and edge cases that Large Language Models frequently exhibit, rather than reiterating fundamental concepts they are already trained on.

Quick Start & Requirements

• Installation: Use npx skills add https://github.com/twostraws/swift-concurrency-agent-skill --skill swift-concurrency-pro.
• Prerequisites: Requires Node.js. If npx is not found, install Node.js via Homebrew (brew install node).
• Usage: Triggered via commands like /swift-concurrency-pro (Claude Code) or $swift-concurrency-pro (Codex), with options for specific review requests.
• Resources: An accompanying YouTube video details Xcode integration.

Highlighted Details

• Covers core Swift concurrency features: async / await, actors, Sendable, task groups, @concurrent, and structured concurrency.
• Specifically designed to address common LLM errors in Swift concurrency.
• Developed by Paul Hudson, author of a book on Swift concurrency and numerous articles on Hacking with Swift.
• Focuses on edge cases, surprising behaviors, and soft deprecations, avoiding basic concepts.

Maintenance & Community

Contributions are welcomed for new checks, improvements, or documentation edits. All work must adhere to the MIT license and the project's Code of Conduct. The project is associated with Hacking with Swift.

Licensing & Compatibility

Licensed under the MIT License, which permits commercial use, modification, distribution, and private use without significant restrictions.

Limitations & Caveats

The skill is intentionally focused on advanced LLM-specific concurrency mistakes and does not aim to cover all Swift concurrency advice or fundamental concepts. It is designed as a specialized tool for AI assistants rather than a general-purpose linter.