SuperAPI AI Directory

code-review

Testing & quality

View repository

Updated 10-16 01:35

About this skill

Code Review Skill

Applicability

  • Platforms: iOS and Android
  • React Native: 0.76+ (New Architecture interop assumed unless a checklist item says otherwise)

When to Use

  • Reviewing a pull request or code change
  • Auditing an existing screen or feature for issues
  • Running a pre-merge quality gate
  • Investigating the root cause of a production bug

This skill is the entry point for a full review. It does not restate every check — each concern lives in its own focused skill. Work through the relevant skills below, then report findings in the output format.

Prerequisites

  • Access to the code being reviewed
  • An understanding of the feature's intended behaviour
  • Knowledge of which data is user-sensitive or business-critical

How to Conduct a Review

Review each concern using its focused skill. Each link resolves to that skill's checklist, examples, and pitfalls — apply those there rather than duplicating them here.

Concern Focused Skill What it covers
Architecture architecture Folder structure, navigation, deep linking, rendering safety, error boundaries, code quality
Performance performance Re-renders, list virtualisation, images, animations, memory and resource cleanup
Accessibility accessibility Labels, roles, touch targets, focus order, colour, dynamic type
Testing testing Unit, component, and E2E coverage; edge cases; test quality
State & Data state-and-data Server state, caching, offline, network transitions, transactions, loading/empty/error states
Security security Secrets, secure storage, transport, PII in logs, deep-link validation, privacy compliance

For each concern: open the focused skill, apply its checklist to the code under review, and collect anything that fails into the output format below. If a concern does not apply to the change, note it as not applicable.

Output Format

Structure every review as:

  1. Summary — overall assessment (solid / needs work / significant issues)
  2. Must-Fix — crashes, data loss, security or compliance issues (block merge)
  3. Should-Fix — performance problems, architecture violations, missing tests
  4. Consider — alternative approaches, future-proofing, style
  5. What's Good — well-implemented patterns worth reinforcing

Pitfalls

  • Reviewing only the happy path — the focused skills exist to make sure offline, error, empty, and accessibility states get checked too.
  • Restating a focused skill's checklist here instead of linking to it causes drift; keep the detail in one place.
  • Skipping the "What's Good" section removes the positive reinforcement that makes reviews land well.