Improve Playwright reliability checks (codex)

[numnx]

🧩 Task: Improve Playwright reliability checks

Automated task PR opened by Code UX for T08.

📋 Summary

Sprint: Sprint 143 · QS: Coverage Observability and CI Reliability Audit · Task: T08

Strengthen Code UX coverage observability and CI reliability through enforceable tests, workflow hardening, structured observability guarantees, security validation, and aligned operational documentation without lowering coverage thresholds.

📝 Full Task Prompt

## Objective
Make Playwright E2E reliability enforceable across clean CI and local runs by validating compiled server startup, `/health` liveness, isolated state, viewport coverage, and artifact behavior.

## Scope
- `playwright.config.ts`
- `tests/e2e/helpers/prepare-app.ts`
- `tests/e2e/example.spec.ts`
- `tests/e2e/accessibility-smoke.spec.ts`
- `tests/e2e/app-release-smoke.spec.ts`
- `tests/e2e/project-setup-release.spec.ts`
- `.github/workflows/playwright.yml`
- `docs/development/testing-and-quality.md`
- `docs/operations/runbook.md`
- Current verification gap: Playwright uses `/health` and temp home, but the audit needs explicit E2E assertions for startup readiness, no fixed local state, viewport coverage, browser cache assumptions, and artifact retention.
- Desired guarantee: E2E runs only against the local compiled app, isolates HOME/USERPROFILE, avoids fixed developer state, checks desktop/mobile viewports where relevant, and uploads useful traces/reports on failure.

## Implementation Requirements
1. Add Playwright tests or helper assertions that call `/health` through `request`, verify the app shell loads from `baseURL`, and never navigate to external websites.
2. Extend fixture helpers to make run IDs deterministic enough for retries while still isolating worker/retry state; assert project fixture names and paths are unique per worker/retry.
3. Add a mobile viewport check to an existing responsive or accessibility spec using Playwright viewport APIs without enabling broad parallelism.
4. Review `playwright.config.ts` artifact settings and ensure trace, screenshot, video, output directory, reporter, `reuseExistingServer: false`, and temp `HOME`/`USERPROFILE` behavior are covered by tests or documented assertions.
5. Harden `.github/workflows/playwright.yml` if needed so build precedes E2E, browser install cannot rely only on cache, artifacts upload on `always()`, and Windows avoids stale `node_modules` cache assumptions.
6. Update `docs/development/testing-and-quality.md` and `docs/operations/runbook.md` with the Playwright startup, health, isolation, viewport, and artifact guarantees.

## Constraints
- Do not require provider credentials, Docker provider startup, worker dispatch, sprint execution, or real project state in E2E tests.
- Keep Playwright `workers: 1` and avoid fixed sleeps or browser-cache-dependent assertions.
- Do not switch readiness from `/health` to `/ready` for clean CI startup.
- Preserve cross-platform behavior for Linux, macOS, and Windows runners.

## Verification
- `pnpm run build`
- `pnpm exec playwright test tests/e2e/example.spec.ts tests/e2e/accessibility-smoke.spec.ts tests/e2e/app-release-smoke.spec.ts tests/e2e/project-setup-release.spec.ts`
- `pnpm exec playwright test`
- `pnpm run lint`
- Success means Playwright starts the compiled server from a clean temp home, passes local health and viewport checks, and CI artifact/cache behavior is documented and reliable.

✅ QA Review

⏳ QA review runs after this PR is opened. Check the task card in the dashboard for the outcome once it lands.

🤖 Model & Provider

Provider	Codex CLI
Model	gpt-5.5

⏱️ Timing

Started	2026-07-06 19:03:04 UTC
Finished	2026-07-06 19:18:09 UTC
Duration	15m 5s

📊 Token Usage

Metric	Value
Input tokens	207,521
Cached input tokens	2,476,672
Output tokens	14,626
Total tokens	2,698,819
Invocations	1
Tool calls	57
Active time	13m 19s

Included usage estimate (subscription/local login): $2.71

🌿 Branch Info

Base: feature/CODUX-143-qs-coverage-observability-and-ci-reliability-audit-3
Head: task/feature-codux-143-qs-coverage-observability-and--t08-codex-mr9l88tm

---

🤖 Generated by Code UX