What does this skill do, and when should you use it?
This skill is for developers upgrading React Native applications or, where relevant, an Expo SDK alongside React Native. It organizes the work into upgrade-path selection, Upgrade Helper template diffs, third-party dependency assessment, React alignment, Expo steps, and post-upgrade verification. Its scope includes package.json, CocoaPods, Gradle, iOS, Android, and common breaking API changes. It fits teams that want a structured upgrade checklist, although the source provides no test suite or verified-platform matrix.
Reads the app's React Native version, queries the latest version, downloads an rn-diff-purge template diff, lists changed files, guides package.json dependency updates and native iOS/Android changes, runs npm install and pod install, and validates Android and iOS builds with React Native and xcodebuild commands.
- A React Native app maintainer moving an application from one 0.x release to a newer release.
- A team reconciling Upgrade Helper template changes in native iOS and Android configuration.
- A project needing third-party dependency compatibility triage and migration planning during an RN upgrade.
- An Expo project maintainer upgrading the Expo SDK together with React Native.
How do you install this skill?
- Before execution, confirm APP_DIR, the version pair, package manager, and target platforms to avoid changing the workspace root or the wrong app.
- The sudo xcode-select and xcodebuild commands have system-level effects and should require explicit user authorization and a recoverable environment.
- curl, npm view, Expo commands, and external references depend on network access; mainland-China reachability and third-party content stability are not guaranteed.
- Do not copy the example App scheme, paths, or versions directly; the project owner must run and record iOS, Android, test, and critical-user-flow verification.
- Shell / CLI
- Network access
- Local filesystem
Node.js/npmcurlgrepCocoaPodsXcode xcodebuildReact Native CLI
The README documents manual installation: run git clone https://github.com/callstackincubator/agent-skills.git, then copy agent-skills/skills/* into a project's .codex/skills/ directory or into user-level ~/.codex/skills/. Restart Codex to recognize newly installed skills. The repository also provides Codex marketplace metadata.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/callstackincubator/agent-skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/upgrading-react-native" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Upgrade this app from React Native 0.76.9 to 0.78.2 and check the iOS, Android, dependency, and Expo impacts.
In a React Native project with the skill installed, use a prompt such as “Upgrade this app from React Native 0.76.9 to 0.78.2 and check the iOS, Android, dependency, and Expo impacts.” The documented example checks the current version, downloads an Upgrade Helper diff, reviews changes, installs dependencies, runs pod install, and validates Android and iOS builds.
What are this skill's strengths and limitations?
- Covers the workflow from upgrade-path selection through build verification.
- Explicitly addresses dependencies, React, Expo SDK, CocoaPods, Gradle, and native platform configuration.
- Includes copyable commands for version checks, diff retrieval, dependency installation, and build validation.
- The source shows no automation script that applies every upgrade change and no test suite.
- The example depends on Node.js/npm, network access, CocoaPods, Xcode, and Android build tooling.
- The source provides no tested operating-system list, supported RN version range, or success-rate data.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| React Native Upgrade Assistant this page | 58 · Recommended | ★ 1.7k | 18d ago | MIT |
| Expo SDK Upgrade Assistant | 50 · Use with care | ★ 2.9k | 21d ago | MIT |
| React Native Brownfield Migration | 53 · Use with care | ★ 1.7k | 18d ago | MIT |
| Expo Web-to-Native Migration | 52 · Use with care | ★ 2.9k | 21d ago | MIT |
| Google Mobile Ads Next-Gen Migration Assistant ✓ Google · Official | 46 · Use with care | ★ 21k | 12d ago | Apache-2.0 |
How did FollowSkills review this skill?
The skill requires target-app confirmation, user confirmation for ambiguous dependency migrations, and recording manual deviations. However, it invokes sudo, network downloads, dependency installation, and native builds without sufficiently documenting data flows, credential handling, permission boundaries, or rollback, so 9 points are deducted.
The workflow is coherent and layered, with version-pair scoping, stop conditions, install guidance, and verification gates. Static evidence is limited by a hard-coded App scheme, some commands lacking failure controls, and no committed tests or standardized abnormal-input feedback; the score stays below the static ceiling.
Triggers, Expo conditions, monorepo boundaries, and APP_DIR scoping are relatively clear. Non-fit ranges and Chinese-language support are not specified, while npm, GitHub, Upgrade Helper, and Apple/Android toolchain access may be unreliable from mainland China, so 4 points are deducted.
The main file provides progressive routing, reference documents, problem mapping, examples, MIT licensing, and stop conditions. It lacks a skill version, changelog, explicit maintenance responsibility, and clear update path; the author field alone does not establish governance, so 5 points are deducted.
It covers dependency, React, Expo, iOS, Android, and verification concerns and offers substantial actionable guidance. The output still requires project-specific merges, human decisions, and bilateral platform validation, and static review cannot prove direct completion, so it receives the static maximum of 7.
The material cites auditable primary sources such as rn-diff-purge, Upgrade Helper, Expo, and React Native, and requests captured verification evidence. It contains no skill-specific test suite, CI coverage, or executed results, so the static maximum of 5 applies.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →