fix: fail CI on test timeout and upload vitest report on failure (closes #706)

[boalambo]

Problem

The test step in .github/workflows/ci.yml wrapped vitest run in a
shell timeout with || exit 0. This caused the CI job to always exit
with code 0 — even when tests failed or the suite timed out — giving
false confidence and allowing broken builds to be merged undetected.

Changes

• Removed the || exit 0 fallback and the shell timeout wrapper
• Added timeout-minutes: 5 directly on the Run Tests step so
  GitHub Actions enforces the limit and fails the job on expiry
• Added --reporter=json --outputFile=vitest-report.json to Vitest
  so a structured report is written during the run
• Added an Upload test report on failure step that uploads
  vitest-report.json as a CI artifact when the job fails

Acceptance Criteria

• A test suite timeout causes the CI job to exit with a non-zero code
• The pipeline blocks the PR merge when tests timeout
• Test output is available as a CI artifact for debugging

Impacted Files

• .github/workflows/ci.yml

Closes #706

[drips-wave]

@boalambo Great news! 🎉 Based on an automated assessment of this PR, the linked Wave issue(s) no longer count against your application limits.

You can now already apply to more issues while waiting for a review of this PR. Keep up the great work! 🚀

Learn more about application limits

[boalambo]

Gm maintainer, looking forward to hearing from you.

[boalambo]

Merge conflicts from #836 are resolved and pushed (build, Type Check/Lint/Validation, and the quality gate are all green ✅).
The test check is failing, and I want to flag why rather than just report it: this is expected, and it's actually confirmation that the fix in this PR works. The old CI step wrapped vitest run in timeout 30s ... || exit 0, silently passing on timeout. This PR removes that fallback and uses timeout-minutes instead, so now the test job is correctly surfacing failures that were previously hidden by the swallowed exit code, not introducing new ones.
I ran the suite locally and traced the failures: they're pre-existing across the repo (missing @/contexts/ThemeContext module, window.matchMedia not mocked in jsdom, pool.test.ts, useSubscriptionClient provider setup, a debounce-timing gap in useDashboardWidgets.test.tsx, etc.) same category already flagged as out-of-scope on #836.
Given the acceptance criteria for #706 is specifically "CI blocks merge on failures instead of hiding them," I think this PR is functionally done, the test check going red here is the intended behavior, not a blocker to fix within this PR's scope. How would you like to proceed, merge this as the CI-visibility fix and track the actual test failures separately, or do you want a broader test-stabilization pass first?

@RUKAYAT-CODER