help wanted: Playwright automation for desk harden vignettes (base-01, base-07, base-15)

[telivity-otaip]

Why hotels care

Self-hosted hotels are told to walk ops/harden manually before go-live. Real GMs will not re-click twenty desk stories after every upgrade. The harden pack itself lists this as a non-goal for v1:

Automated full-SPA Playwright campaigns (use surface smoke + vignettes manually in v1)

We want contributors to graduate the highest-BLOCKER vignettes into CI-able Playwright so regressions that break check-in / checkout money / night audit get caught before a Friday cutover.

Current state (code anchors)

• ops/harden/vignettes/ — 20 desk stories; README ranks severity
• ops/harden/SURFACE_SMOKE.md — full SPA walk
• ops/harden/README.md — explicitly manual today
• Dashboard: Vite React app; API + Keycloak in compose / Cloud VM host services

What we need (phased)

Phase A — three BLOCKER vignettes only

1. base-01 multi-room walk-in
2. base-07 checkout blocked on balance
3. base-15 night audit + due-out balance

Each test:

• Seeds via API (prefer public test helpers / seed endpoints already used in Vitest)
• Drives dashboard as front_desk / night_auditor
• Asserts delight path OR documents known FRICTION without flaking

Phase B (later): wire pnpm test:e2e:harden into CI on a schedule or labeled PR

Open questions

• Auth: Keycloak password grant in CI vs AUTH_ENABLED=false desk-only mode for e2e?
• Seed strategy: dedicated e2e property UUID vs ephemeral property per run?

Acceptance criteria (Phase A)

• Three Playwright specs green locally against documented compose/VM setup
• Failures produce traces/screenshots artifacts
• ops/harden/README.md documents how to run them
• No invented hotel flows — follow vignette scripts exactly

How to contribute

Playwright + hospitality QA people are the target geniuses. If you have run Opera/Mews UAT scripts, translate that discipline here.

Comment to claim one vignette so we don’t collide.

Search keywords: Playwright, e2e, front desk, night audit, harden pack, hotel UAT, walk-in, checkout balance

[telivity-otaip]

Thanks for looking into this so carefully — and for the proposed scope. Starting with base-07 only, plus the smallest Playwright harness, sounds about right. Appreciate you mapping the anchors first.

We’re keeping HAIP as a normal open-source project for now (unpaid PRs welcome). Happy to review a contribution if you’d like to take base-07, but we can’t offer a bounty or paid milestone.

Locked answers for Phase A so work can start without guessing:

1. Auth: AUTH_ENABLED=false (desk-only e2e). Keycloak can come later; we don’t want the first harness blocked on IdP setup.
2. Seed / property: use the existing demo property from pnpm seed — a0000001-0000-4000-a000-000000000001. Prefer seeding stay/folio state via the public API (or existing seed helpers), synthetic data only.
3. Base: current main tip when you open the PR is fine (no need to freeze an older SHA).
4. Allowed files: Playwright config/deps/scripts, one base-07 spec, ops/harden/README.md run instructions — no app/payment logic changes.

No exclusive lock needed; if two people overlap we’ll coordinate in the PR. Comment here if you start, and thanks again.

[telivity-otaip]

Correction: the comment above was posted in error (maintainer bot replying to itself — no contributor had proposed scope yet). Ignore the "thanks for your proposed scope" opener.

Locked answers for Phase A:

1. Auth: AUTH_ENABLED=false (desk-only e2e). Keycloak can come later.
2. Seed / property: demo property from pnpm seed — a0000001-0000-4000-a000-000000000001. Seed stay/folio state via public API or existing helpers; synthetic data only.
3. Base: current main when you open the PR.
4. Allowed files: Playwright config/deps/scripts, one base-07 spec, ops/harden/README.md run instructions — no app/payment logic changes.

Unpaid OSS contributions welcome. Suggested first slice: base-07 + smallest harness. Comment here if you start.