0001. Scope, audience and the two-track plan

• Status: Accepted
• Date: 2026-10-10
• Questions: A1–A7

Context

Existing templates each solve part of the problem (see docs/research.md). We want one template that keeps their strengths and fixes their gaps, for both the bare React Native CLI and Expo.

Decision

• Two tracks. Track A (this repo) is a bare React Native CLI template. Track B ports the same src/ to Expo with Continuous Native Generation after Track A ships. Native and vendor capabilities sit behind ports so Track B swaps adapters, not features.
• Audience (A1): everyone. A core that suits solo developers, agencies and product teams, with extras opt-in.
• Distribution (A2): both a GitHub template repository and an npm template for npx @react-native-community/cli init MyApp --template …, built from the same source.
• Opinionation (A3): fixed core + opt-in modules. Core is always on; modules are added by recipe.
• Example code (A4): full demo (auth, feed, settings, component showroom) that a script can remove.
• Backend (A5): REST + OpenAPI, with an MSW mock server so the app runs with no backend.
• License (A7): MIT.
• Name (A6): working name rn-template until a final name is chosen. Renaming is one pnpm bootstrap run. Superseded by 0015: Bareship.

Consequences

• The demo features must stay removable: scripts/strip-examples.ts is part of the deliverable. Built 2026-10-12: it removes the feed, posts mocks and showroom (marked // @example in the code) and keeps sign-in, onboarding and settings, which most apps need. scripts/ci/check-strip-examples.sh checks the stripped app.
• Ports add a small indirection everywhere a native library is used. In exchange, Track B and vendor swaps stay cheap.

Alternatives considered

• Expo first: faster, but hides the native decisions this template wants to make explicit.
• Pick-everything CLI: maximum flexibility, but the maintenance matrix explodes.