What does this skill do, and when should you use it?
This skill is for teams that have already chosen a brownfield migration strategy. It routes the work through either an Expo or bare React Native track, then covers integration setup, artifact packaging, and host-app integration. The workflow includes iOS XCFramework and Android AAR outputs, followed by phased adoption by feature or screen. It is a strong fit for existing native apps seeking gradual React Native adoption rather than a full rewrite.
Inspects project files to determine whether React Native, Expo configuration, and native folders are present; routes the task to either the Expo or bare React Native track; guides initial setup with @callstack/react-native-brownfield; documents generation of iOS XCFramework and Android AAR artifacts; guides integration into native host apps; and uses agent-device for startup/runtime verification through app launch, navigation, snapshots or screenshots, and device evidence.
- A native iOS team has approved incremental migration and needs to embed its first React Native surface.
- A native Android team needs to generate, publish, and consume an AAR from a React Native source app.
- An existing Expo project needs Expo plugin setup and brownfield packaging readiness.
- A bare React Native project needs XCFramework or AAR generation and integration into a native host app.
How do you install this skill?
- Before running npx, npm install, CocoaPods, Gradle, or publishing commands, confirm dependency versions, provenance, permissions, and rollback state.
- The references require agent-device but do not define its installation source, permission boundaries, or a complete alternative when no device is available.
- Several key paths depend on omitted reference files and external documentation, so API and current-version compatibility cannot be confirmed from this material alone.
- No fallback is documented for inaccessible documentation, package registries, or tooling from mainland-China networks.
- Shell / CLI
- Network access
- Local filesystem
@callstack/react-native-brownfieldagent-device
The README documents repository-level installation: clone the repository, then copy agent-skills/skills/* into a project .codex/skills/ directory or into ~/.codex/skills/, and restart Codex. The source does not provide a dedicated command for installing only this skill.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/callstackincubator/agent-skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/react-native-brownfield-migration" ~/.claude/skills/
rm -rf "$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:
- Integrate one React Native screen into our existing native iOS app and produce the XCFramework.
Invoke it after the migration decision has been made, for example: “Integrate one React Native screen into our existing native iOS app and produce the XCFramework.” It first determines whether the project follows Expo or bare React Native. If the path remains unclear, answer the required disambiguation question. Use exactly one track per task.
What are this skill's strengths and limitations?
- Covers both Expo and bare React Native brownfield tracks.
- Includes setup, XCFramework/AAR packaging, and native host integration guidance.
- Requires packaging validation before host integration.
- Supports phased rollout by feature or screen.
- It is not intended to decide whether brownfield migration is appropriate in the first place.
- The main file routes to several reference files whose detailed commands are not included in the supplied source.
- The source provides no test-suite results, platform coverage report, or evidence of successful builds.
- Device verification depends on agent-device being available or separately enabled.
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 Brownfield Migration this page | 53 · Use with care | ★ 1.7k | 19d ago | MIT |
| Expo Brownfield Integrator | 58 · Recommended | ★ 2.9k | 22d ago | MIT |
| React Native Migration Assessor | 57 · Use with care | ★ 1.7k | 19d ago | MIT |
| Expo Web-to-Native Migration | 52 · Use with care | ★ 2.9k | 22d ago | MIT |
| Expo Native UI Builder | 59 · Recommended | ★ 2.9k | 22d ago | MIT |
The skill explicitly routes between Expo and bare React Native. Expo is selected when Expo is present or its plugin workflow is requested; bare React Native is selected when native folders and a direct React Native CLI workflow are used without Expo path requirements.
How did FollowSkills review this skill?
The material shows path-selection gates, placeholder resolution, staged validation, stop conditions, host isolation through a facade, and trusted-path guidance for installing missing verification tooling. It lacks explicit user confirmation, detailed permission boundaries, sensitive-data and dependency-supply-chain guidance, rollback procedures, and external-effect disclosure, so points are deducted.
The workflow has clear prerequisites, checklists, failure guidance, and stop conditions. However, the static material contains no test suite or execution results, several referenced key files are not provided, and API, version-compatibility, and abnormal-input handling remain uncertain. The score is therefore conservative and respects the static-review cap.
The intended audience, Expo and bare scenarios, path-selection gate, and non-fit migration-decision scenario are reasonably clear. Non-fit boundaries, version/environment matrices, Chinese-language support, and mainland-China network reachability are not addressed, and the workflow depends on external documentation, so points are deducted.
The evidence includes YAML metadata, an MIT license, layered references, quick-reference routing, related skills, common pitfalls, and troubleshooting sections. It does not clearly provide a version, changelog, maintenance owner, or update policy, and some linked targets are absent from the supplied material, so points are deducted.
The skill covers the core setup, packaging, publishing, host-integration, and runtime-verification flow, with commands and snippets that are directly useful as references. However, results depend heavily on project structure and tool versions, no execution evidence is supplied, and substantial manual adaptation and review remain necessary. The static-review ceiling also limits the score.
The skill lists pinned-style canonical documentation locations, command success criteria, and runtime verification requirements. It provides no committed tests, CI coverage, or third-party execution records; key claims remain primarily author-authored documentation, so only limited points are justified.
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 →