fix: atomic daily review XP cap check

[yush-1018]

This PR resolves #400 by enforcing the daily help review XP cap check atomically inside a database transaction rather than checking the count first in a separate non-atomic query.

Changes Made

• Atomic Transaction Check: Wrapped the count increment in \xp_daily_usage\ and the insertion in \xp_events\ inside a database transaction in \src/lib/xp/events.ts. If the cap is exceeded, we throw \daily_review_cap_reached\ to roll back. If a duplicate event occurs, we call \ x.rollback()\ to undo the daily count increment.
• Webhook Integration: Updated the GitHub review event handler (\processReviewEvent) to use the new atomic check and gracefully catch the limit error.
• Unit Tests: Added coverage for rollback on limit breach and duplicate inserts inside \src/lib/xp/events.test.ts.

Verification

• 
  pm run typecheck\ passes cleanly.
• 
  px vitest run src/lib/xp/events.test.ts\ passes with 13/13 tests green.
• 
  px vitest run src/inngest/functions/process-review-event.test.ts\ passes with 5/5 tests green.

[vercel]

@yush-1018 is attempting to deploy a commit to the codersogs-3057's projects Team on Vercel.

A member of the Team first needs to authorize it.

[Ayush-Patel-56]

this catch swallows any transaction error that isn't the cap message and just treats it as inserted = false - a real DB failure here would silently lose the reviewer's XP with no log or retry, since process-review-event's catch never gets a chance to see it. can you narrow this to only catch the duplicate/rollback case and rethrow everything else?

[jakharmonika364]

Looks good. Thanks!