Production Rescue · 1–2 weeks

Ship the fixes, not another report

We join the repository and fix the foundation. The UI stays. The security model, the queries, and the deploy path get rewritten until they hold up.

Critical bug fixes on the paths customers actually hit
Authentication, authorization, and RLS rewrites
Database indexes, connection pooling, and API timeouts
Idempotent webhooks and rate limits
Deployment verification on staging, then production
Tests around the code we change, plus a warranty on that scope

This is the engagement for Lovable production issues, Bolt.new apps that fall over under real usage, and Cursor or Claude Code codebases that need AI-generated app debugging with pull requests instead of advice. New product features are out of scope.