Feat

Feature Workflow

Conventions this workflow depends on (read the ones your change touches before writing code):

• CorePlatformModules.md — required before touching packages/core (platform-split files, handwritten .d.ts).
• CodeComments.md — comment rules for all TypeScript edits.
• WritingUnitTests.md — test expectations and environment constraints.
• CONTRIBUTING.md — commit message format.

Phase 1: Understand

• If given a GitHub issue number, fetch it: gh issue view <number> --json title,body,labels,comments. Otherwise treat the input as a free-text description; if empty, ask for one.
• Summarize the user-facing goal, acceptance criteria, and edge cases before touching code.

Phase 2: Ground in the codebase

• Find an existing module or feature similar to what you are building and use it as the pattern to follow — reference it by path in your plan.
• Map the touch surface: which package, which modules, and for packages/core whether the change lands in -common.ts, both platform files, or all three — plus the neighboring .d.ts.

Phase 3: Plan

Present a short plan before implementing: files to create/edit with a one-liner each, the test strategy, and any public API changes (these require .d.ts updates). Prefer the simplest solution that reuses existing patterns.

Phase 4: Implement

• Follow the referenced conventions. Keep .ios.ts / .android.ts in parity; never leave one side diverged.
• Public API changes update the neighboring .d.ts in the same change, with JSDoc in the .d.ts.

Phase 5: Test

• Add or extend colocated *.spec.ts specs for shared-logic changes.
• Run the focused tests first (npx nx run core:test -t 'Name'), then the full package target (npx nx run core:test) — it must be green.
• Behavior that needs the real native runtime is covered in apps/automated, not unit tests; note it for the PR's manual test scenarios instead.

Phase 6: Finish

• Format: npx nx format:write.
• Commit with the conventional format, e.g. feat(core): <subject>.
• Open the PR with gh, following the PR template: reference the issue and include tests for the change.