export default ios-testing-skill

name	ios-testing-skill
description	Skill describing modern iOS testing patterns: Swift Testing vs XCTest, XCUITest best practices, and OS 2.2 expectations for coverage and flakiness.

iOS Testing Skill – Swift Testing, XCTest, XCUITest

This skill centralizes iOS testing guidance so that testing agents and gates can stay lean while following consistent patterns.

It is used by:

• ios-testing-specialist
• ios-ui-testing-specialist
• ios-verification
• ios-architect / ios-builder when planning or implementing tests.

Core Concepts

• Swift Testing vs XCTest

  • Prefer Swift Testing for Swift 6+ targets where available.
  • Maintain parity when migrating from XCTest; do not regress coverage.
  • Use @Test and #expect/#require with clear naming and tagging.

• Unit vs Integration vs UI

  • Unit tests: small, isolated; no network/DB unless explicitly integration.
  • Integration tests: limited external dependencies; deterministic configs.
  • UI tests (XCUITest): page objects, accessibility identifiers, async-safe waits.

• Testing SwiftData Models

  • Use ModelContainer with isStoredInMemoryOnly: true for tests
  • Test @Relationship cascade deletes verify dependent data is removed
  • Verify FetchDescriptor queries return expected results
  • Use @MainActor for SwiftData operations in async tests
  • Test unique constraints (@Attribute(.unique)) enforce uniqueness

• Testing SwiftUI Atoms

  • Use AtomTestContext for isolated atom testing
  • Override atoms with .override() for dependency injection
  • Test computed atoms (ValueAtom) derive correct values from dependencies
  • Verify StateAtom mutations trigger dependent atom updates
  • Test atom dependency graphs with graphDescription()

Context7 Libraries

Agents using this skill MAY consult:

• os2-ios-testing

  • Swift Testing patterns and examples,
  • XCTest/XCUITest structure and migration guidance,
  • Flakiness mitigation strategies and CI integration tips.

• /websites/developer_apple_swiftdata (via Context7)

  • In-memory ModelContainer patterns for testing
  • Testing @Relationship cascade rules
  • Async/await patterns with @MainActor
  • FetchDescriptor query testing

• /ra1028/swiftui-atom-properties (via Context7)

  • AtomTestContext for isolated state testing
  • Mocking atoms with .override() for dependency injection
  • Testing atom dependency graphs
  • Verifying computed atoms derive correct values

Usage Pattern

1. When designing a test strategy:

   • Decide which layers to cover (unit/integration/UI),
   • Choose Swift Testing vs XCTest based on project/tooling,
   • Define critical paths, error cases, and edge cases up front.

2. When implementing tests:

   • Keep tests deterministic and fast by default,
   • Use tags/traits for slow/critical suites,
   • Structure UI tests around page objects and accessibility identifiers.

3. When testing data persistence (SwiftData):

   • Always use in-memory ModelContainer in tests
   • Test cascade deletes remove dependent data
   • Verify unique constraints throw on duplicates
   • Test relationship integrity (one-to-many, many-to-one)
   • Use @MainActor for async SwiftData tests

4. When testing state management (Atoms):

   • Use AtomTestContext to isolate atom behavior
   • Override dependency atoms with .override() for mocking
   • Test computed atoms recalculate when dependencies change
   • Verify StateAtom mutations propagate to watchers
   • Test atom cleanup when watchers are removed

5. When acting as a gate (verification/quality):

   • Look for coverage of happy paths + failure/edge states,
   • Call out flakiness risks (sleep-based waits, environment dependence),
   • Recommend incremental improvements, not full rewrites, unless required.
   • Verify SwiftData tests use in-memory containers
   • Verify Atom tests use AtomTestContext, not real app state

This skill ensures iOS agents talk about testing in a consistent way and know when to consult context7 for deeper patterns and examples.