菜单
Audit entry point for reviewing React Native code. Routes each concern to a focused skill and defines the review output format. Use when reviewing a PR, auditing a screen, or running a pre-merge quality gate.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Review React Native code for screen-reader support, labels and roles, touch targets, focus order, colour contrast, and dynamic type. Use when building UI, reviewing components, or auditing a screen for accessibility.
Review and structure React Native features for correct folder layout, navigation, deep linking, error boundaries, rendering safety, and code quality. Use when designing a new feature, reviewing structure, or isolating crashes.
Baseline React Native project conventions for structure, TypeScript, naming, file size, and hygiene. Use when setting up a project, writing new code, or reviewing for consistency.
Non-negotiable React Native rules that prevent crashes, data loss, security breaches, and compliance violations. Use on every change as a baseline safety check, regardless of the task.
Review React Native code for rendering performance, list virtualisation, memoisation, image handling, animations, and memory/resource cleanup. Use when a screen feels janky, before merging UI-heavy code, or when profiling.
Review React Native code for secret handling, secure storage, transport security, PII in logs, deep-link validation, input validation, and privacy compliance. Use when handling auth, payments, personal data, or before a security review.
| name | code-review |
| description | Audit entry point for reviewing React Native code. Routes each concern to a focused skill and defines the review output format. Use when reviewing a PR, auditing a screen, or running a pre-merge quality gate. |
| version | 2.0.1 |
| platforms | ["ios","android"] |
| react-native-version | 0.76+ |
| tags | ["react-native","code-review","audit"] |
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.
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.
Structure every review as:
