chore: bump @friggframework/* deps to 2.0.0-next.82

[d-klotz]

Summary

Bump @friggframework/* deps from 2.0.0-next.81 → 2.0.0-next.82. Pulls in Frigg PR #582 (fix(core): explicitly persist tokens on OAuth2 re-auth).

Why

Diagnosed against Pipedrive in prod (user 10, integrations 12378 + 12379, both currently ERROR):

• User reinstalls the integration → OAuth callback runs against Pipedrive → Pipedrive issues fresh access + refresh tokens (and retires the old refresh_token).
• Frigg's ProcessAuthorizationCallback runs getToken but the DLGT_TOKEN_UPDATE notification chain that's supposed to persist the new tokens silently no-ops in prod. The user-visible /api/authorize returns success.
• The DB still holds the 3-month-old tokens, which Pipedrive has now invalidated.
• Downstream webhook setup then 401s, integration flips to ERROR. User has no clear signal of what went wrong.

Confirmed by decrypting credential 936: both the access_token (401 invalid) and the now-stale refresh_token (invalid_grant) come from Feb 5, never replaced today despite two reinstalls.

What's in next.82

• Bootstraps the Module with the existing entity on re-auth so the API requester can use prior tokens as a fallback (replaces a long-standing // todo left as entity = null).
• Belt-and-suspenders explicit upsertCredential after getToken on the OAuth2 path. The notification chain stays in place; the explicit call ensures persistence even when the chain doesn't fire.
• Diagnostic logging at POST /api/authorize entry, after getToken, after persistence, and after restore — so future re-auth bugs surface in CloudWatch instead of running silently.
• restoreIntegrationsForEntity continues to flip linked ERROR/DISABLED integrations back to ENABLED on successful re-auth (existing behavior).

Test plan

• npm install resolves cleanly with the new versions
• Deploy to dev; have a user re-install Pipedrive; confirm CloudWatch shows the new [Frigg] processAuthorizationCallback ... log lines
• Verify the credential's updatedAt matches the OAuth callback time (not a later auth-flip)
• Re-auth on integration 12379 in prod; expect status to flip back to ENABLED and sync to resume

🤖 Generated with Claude Code