automatelisted

# automate — Gherkin test book → native Playwright (POM-as-fixtures) Turns `.feature` scenarios into executable Playwright tests, following the reference proven in `examples/medibook/` (26 tests, 32 executions — the e2e suite runs on two device profiles). Two architectural choices are fixed, not renegotiated per run: - **Native Playwright, no Cucumber layer.** The test book stays Gherkin; each generated test carries its scenario's stable ID instead of re-parsing the `.feature` at runtime. - **Page objects exposed as Playwright fixtures**, not a classic inheritance-based POM. ## Prerequisites - A QAIA test book (`.feature` files with stable `@QAIA-<US-ID>-<NNN>` tags) — from `qaia-core:testbook-generate`. - A running target app the user designates (URL). Automation needs a real environment; say so if none is provided. ## Rules (non negotiable — from the medibook reference) - **POM as fixtures** — one page object per screen under `pages/`, selectors by role or `data-testid` only, **no assertions inside page objects**: assertions live in tests. Page objects are exposed as Playwright fixtures so each test gets fresh instances. - **Traceability** — every generated test title carries its source scenario ID and AC tag (`@QAIA-US-001-003 @AC5`), the same IDs the test book uses. Requirement → scenario → automated test stays one continuous chain. - **Atomic preconditions** — each test seeds its own state declaratively (API seeding or fixtures), never a UI-chained set