agentleFS
maroffomaroffo/claude-forge/skills/apple-swift/SKILL.md

apple-swift

Skill17 starsChanged 51 days agoMITView on GitHub

Apple platform development with Swift, SwiftUI, async/await, and performance. Use when working with .swift files, Package.swift, Xcode projects, or building for iOS/macOS/watchOS/visionOS.

The file

Files
SKILL.md

ABOUTME: Apple platform guide, Swift, SwiftUI, concurrency, testing, performance

ABOUTME: Conventions, strict concurrency rules, MVVM + DI patterns, review checklists

Apple Platform Development

Quick Reference

# Build
xcodebuild -scheme MyApp -sdk iphoneos build
xcodebuild -scheme MyApp -sdk macosx build

# Tests (pick <booted device> from: xcrun simctl list devices | grep Booted)
xcodebuild test -scheme MyApp -destination 'platform=iOS Simulator,name=<booted device>'

# SwiftLint / SPM
swiftlint lint [--fix]
swift build && swift test && swift package resolve

# ast-grep
sg --pattern '@Observable final class $NAME { $$$ }' --lang swift
sg --pattern '@MainActor' --lang swift

See also: _AST_GREP.md, _PATTERNS.md, source-control, references/


Version (determine, don't assume)

See ../_LANG_COMMON.md. Fetch the truth:

swift --version                                 # local toolchain
xcodebuild -version                             # Xcode version
cat .swift-version 2>/dev/null                  # pinned toolchain (if present)
grep -E 'swift-tools-version' Package.swift     # SPM minimum
curl -s https://api.github.com/repos/swiftlang/swift/releases/latest | jq -r .tag_name  # latest upstream

Pre-Commit Verification (MANDATORY)

make check && make test-e2e must pass (enforced by the pre-commit-gate hook; see ../_LANG_COMMON.md). What make check expands to for an Apple project:

swift build                                     # compilation check
swift test                                      # unit tests
swiftlint lint --strict                         # lint (fail on warnings)
swift-format lint --recursive Sources Tests     # formatting check
xcodebuild -scheme MyApp test \
  -destination 'platform=iOS Simulator,name=<booted device>'   # platform tests

Pick <booted device> from xcrun simctl list devices | grep Booted (do not hardcode a simulator model).


SwiftUI

Property Wrappers

WrapperUse
@StateView-owned values, @Observable
@BindingTwo-way to parent
@BindableTwo-way to @Observable props
@EnvironmentSystem/app values
@StateObjectView-owned ObservableObject (legacy)

View property order: @Environment, let, @State/@Binding, computed, init, body, private methods

View sizing: <100 lines = single view; 100-200 = extract subviews; >200 = multiple files; business logic = @Observable VM; network/DB = repository pattern

For NavigationStack, SwiftData, MVVM, and DI patterns, see references/swiftui-patterns.md.


Concurrency

Common Fixes

ErrorFix
Main actor isolationAdd @MainActor to class/func
Non-isolated accessMark nonisolated
Sendable violationAdd @unchecked Sendable or fix
Protocol asyncRequire async in protocol
Closure captureUse @Sendable closure

When to use what

Use CaseChoice
One-shot networkasync/await
Parallel fetchesasync let / TaskGroup
Real-time streamsCombine / AsyncStream
UI events, debounceCombine

UI updates run on @MainActor, cross-actor value types are Sendable, and long-running work checks Task.isCancelled: under Swift 6 strict concurrency the first two are compile errors, the third is wasted work that nothing reports.

For MainActor, parallel execution, and actor patterns, see references/concurrency-patterns.md.


Architecture

MVVM with @Observable: @Observable @MainActor final class VM with private(set) properties, injected service protocols, async load methods with defer { isLoading = false }.

DI: Protocol-based with EnvironmentKey for SwiftUI injection.

Coordinator pattern: For flow-heavy apps, lift navigation state out of views into a Coordinator type owning a NavigationPath (or route enum) and exposing intents.

Testing (Swift Testing): @Suite, @Test, #expect. Parameterized with arguments:. Mock via protocol injection. XCTest still valid where needed (UI tests, legacy targets).

For full code examples, see references/swiftui-patterns.md and references/swift6-patterns.md.


Review Checklists

Concurrency: MainActor for UI, Sendable for cross-actor data, task cancellation handled, no data races

SwiftUI: @Observable over ObservableObject where available, NavigationStack over NavigationView, .task over .onAppear + Task, LazyVStack for long lists

Critical: Force unwrap without safety, UI updates off MainActor, data races, retain cycles

Major: Legacy ObservableObject when @Observable fits, NavigationView in new code


Detailed References

  • references/swift6-patterns.md - Strict concurrency migration, Sendable, actors, macros
  • references/swiftui-patterns.md - NavigationStack, SwiftData, MVVM, dependency injection
  • references/concurrency-patterns.md - async/await, TaskGroup, MainActor, actors, AsyncStream
  • references/performance.md - Optimization, Instruments profiling, memory management

Official: Swift | SwiftUI | SwiftData

Libraries: TCA | Snapshot Testing | SwiftLint

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.