ftk-ci

Files (1)

• SKILL.md

SKILL.md (first lines)

Generate `.github/workflows/flutter.yml` fitted to THIS project — read `pubspec.yaml` (Flutter SDK constraint, flavors/dart-defines, build_runner) before writing.

## Baseline workflow

```yaml
name: flutter-ci
on:
  push:
    branches: [master, main]
  pull_request:

jobs:
  check:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: subosito/flutter-action@v2
        with:
          channel: stable
          cache: true
      - run: flutter pub get
      - run: flutter analyze
      - run: flutter test --exclude-tags golden
      - run: flutter build apk --debug
      - uses: actions/upload-artifact@v4
        with:
          name: debug-apk
          path: build/app/outputs/flutter-apk/app-debug.apk
```

## Adjustments

- Pin the Flutter version to the project's constraint (`flutter-version:`) when the team needs reproducibility over freshness.
- build_runner project → insert `dart run build_runner build --delete-conflicting-outputs` before analyze.
- Flavors → build the dev flavor for the smoke build (`--flavor dev --dart-define=ENV=dev` per project convention).
- **Golden tests are excluded by default** (`--exclude-tags golden`): goldens are generated on the developer machine and are platform-sensitive. Verify golden tests actually carry `tags: 'golden'` (rule in `knowledge/testing.md`) — if untagged goldens exist, tag them as part of this setup. Offer the alternative (regenerate goldens on a Linux runner and run them in CI) only if the user asks for CI-enforced goldens.
- iOS build needs a macOS runner + signing secrets — scope it separately; do not add it silently.

Finish: commit nothing yourself unless asked; run `flutter analyze && flutter test --exclude-tags golden` locally once to prove the commands the workflow uses are green.

From adiwirya/flutter-tools-kit, MIT.

More from adiwirya/flutter-tools-kit