What I have published
The method is open before anyone pays for it. Each of these shows a piece of how a review is done, or what one produces.
2026-08-28
Why you should audit your AI-built app before you rebuild it
Every development agency will tell you to throw away your Cursor or Lovable prototype and quote a $15,000 rewrite. Why a 1-week bounded audit is the safer, cheaper first step.
2026-08-28
The pre-launch disaster drill: migrations, rollback scripts, and PITR for Supabase
Having automated cloud backups is not the same as disaster recovery. How to test Point-in-Time Recovery (PITR), write backward-compatible PostgreSQL migrations, and prepare rollback scripts before your beta launch.
2026-08-28
How Stripe webhooks silently break in AI-generated apps (and how to fix them)
Why AI-coded payment handlers often fail in production: raw body signature parsing in Next.js, missing idempotency guards, and duplicate subscription provisioning on network retries.
2026-08-28
How to take over and stabilize an AI-built codebase without losing momentum
You built an MVP with Claude Code, Cursor, or an outsourced contractor who used AI tools. Now what? A 3-step technical playbook for taking ownership, locking platform boundaries, and shipping features safely.
2026-08-28
Why AI-built apps leak data: the invisible Supabase RLS boundary
Hiding buttons in React is not security. How AI coding tools accidentally leave Supabase PostgREST endpoints wide open, and how to verify multi-tenant isolation with two test accounts.
2026-05-20
7 things to check before you launch an AI-built MVP
A checklist for what to check before you put an AI-coded MVP in front of customers: auth, data ownership, secrets, payments, deployment, failure states, and maintainability.
2026-05-05
What a Remodeling Sprint is: turning a prototype into a product
A service explainer for turning an AI-built prototype into a more reliable product by stabilizing one critical flow before adding more features.