feat: add Acid Tools footer badge

[andrewzolotukhin]

Original request

Add an Acid Tools badge to the public footer using the supplied light and dark hosted badge images and link, following the PR workflow without creating a blog post.

Summary

Adds an accessible, theme-aware Acid Tools directory badge to the public xpenser footer and verifies the integration at component and browser-test levels.

What changed

• Added an AcidToolsBadge to the public footer badge collection.
• Uses https://acidtools.com/assets/images/badge.png in light mode and https://acidtools.com/assets/images/badge-dark.png in dark mode.
• Preserves the supplied Acid Tools alt text, 54px height, lazy loading, target URL, and new-tab behavior while adding the repository-standard focus styling and safe link relation.
• Added component assertions for link attributes, both hosted image variants, height, lazy loading, and theme classes.
• Added Playwright coverage for light/dark visibility and the supplied badge contract.
• After the first CI run exposed that the off-screen lazy image has a zero-width box before loading, updated the e2e flow to scroll the badge into view before asserting visibility. Production markup was unchanged.

Reasoning

The implementation follows the existing hosted footer-badge pattern so the new badge participates in the same responsive wrapping, focus treatment, and dark-mode switching. The supplied markup does not define a width, so the implementation intentionally leaves width intrinsic rather than inventing a dimension. The e2e flow triggers lazy loading as a user would before checking the rendered theme variant.

Blog post

Skipped by explicit user request. This badge-only addition is too small to justify a standalone release post.

Screenshots / preview evidence

Preview: https://xpenser-pr-078.cleverbrush.com

Manual QA exercised the footer at 390 × 844 and 1440 × 900 in light and dark themes, with light/dark mobile and desktop screenshots captured during the session. Verified:

• The light and dark Acid Tools assets both return HTTP 200 and render at their natural 175 × 54 dimensions.
• Only the correct theme variant is displayed.
• The exact destination, target="_blank", rel="noopener noreferrer", alt text, height, and lazy-loading attributes are present.
• The badge container and page have no horizontal overflow at either viewport.
• Keyboard focus shows the repository-standard two-pixel focus ring.
• No browser errors or console output occurred during the badge QA flow.
• External destination navigation was not exercised because the browser safety layer blocked sending the preview referrer off-site; the DOM href was verified exactly.

The passing Playwright e2e job provides stable CI evidence for the updated public homepage flow.

Validation

• npm test -- apps/web/components/landing-page.test.tsx
• npm run lint
• npm run typecheck
• npm test (95 files, 520 tests)
• PLAYWRIGHT_BASE_URL=https://xpenser-pr-078.cleverbrush.com npm run test:e2e -- tests/e2e/public-auth.spec.ts (6 tests)
• Lint and test
• Deploy PR environment
• Playwright e2e
• Preview QA in light and dark modes at mobile and desktop sizes
• SigNoz traces, logs, and metrics explicitly checked; recorded as an infrastructure telemetry skip because xpenser-web-pr-78 never appeared as a trace or log resource and the queried metric series were empty
• PR-ready notification attempted; skipped because the configured default notifier at 127.0.0.1:8787 was unavailable

Telemetry

For the post-QA xpenser-web-pr-78 window:

• Trace queries returned no rows, including the explicit error-trace query.
• ERROR/FATAL log queries returned no rows.
• http.server.duration.count and nodejs.eventloop.utilization returned no matching metric series.
• SigNoz resource-value discovery lists xpenser preview services through xpenser-web-pr-77 but does not contain PR 78 for traces or logs, even after an ingestion wait and recheck.
• The successful deployment confirms the PR OTLP endpoint and repository-standard service-name configuration were supplied, and the preview remained healthy under manual and CI traffic. Because this footer-only diff does not touch telemetry configuration, missing PR 78 ingestion is documented as an infrastructure skip rather than a badge blocker.

Screenshots / Preview

See the preview evidence above and the passing Playwright job for stable browser-test evidence.

Checklist

• I kept the change focused.
• I updated docs or tests where needed (tests updated; blog intentionally skipped as noted above).
• I checked for secrets, local env files, and generated build output.
• API changes keep contracts, endpoint metadata, and handlers aligned (not applicable; no API changes).