SuperAPI AI Directory

swift-code-review

Testing & quality

View repository

Updated 10-16 05:48

About this skill

Swift Code Review

Lead with actionable findings ordered by severity. A review is not a narration of the diff.

Evidence workflow

  1. Read the change, surrounding contracts, callers, manifest, platform conditions, and relevant tests.
  2. Reproduce or statically trace each suspected defect.
  3. Check optionals, value/reference semantics, ARC cycles, errors, Codable shape, public ABI/API, actor isolation, Sendable crossings, cancellation, resources, and C/Objective-C boundaries.
  4. Confirm tests would fail for plausible defects and use production paths.
  5. Report location, trigger, impact, and the smallest credible remedy.

Severity

  • Critical: data loss, exploitable security defect, broad outage, or irreversible compatibility break.
  • High: common wrong result, race/leak, broken public contract, or migration parity failure.
  • Medium: bounded correctness, resilience, performance, or maintainability risk.
  • Low: concrete improvement with limited operational impact.

Swift-specific checks

  • Force unwrap/cast, implicitly unwrapped Objective-C values, lifetime and retain-cycle errors.
  • Incorrect struct/class choice, Hashable/Equatable drift, mutation through shared storage.
  • Codable key/default/null shape changes and error erasure.
  • Unstructured/detached task leaks, actor reentrancy assumptions, unchecked Sendable, swallowed cancellation.
  • SwiftPM target/resource errors, unsafe flags, unsupported platform imports, and secret exposure.

If no actionable defect is found, say so and state remaining verification gaps.