AI app error fix index — 43 common AI app problems

The symptom-to-fix index for AI-built apps. Forty-three exact-match errors drawn from rescue work across Lovable, Bolt, Cursor, Claude Code, v0, Base44, and Replit Agent — now covering Turbopack, Clerk, NextAuth, Prisma pooling, Supabase Realtime and Storage, upstream rate limits, and three AI-builder coding-loop failures (token spirals, regression loops, context loss). Each page maps the error string to the root cause, the code-level fix, and a regression test you can merge.

The Afterbuild Labs /fix index is grouped into four mid-tier topical hubs — payments, auth, deploy, and database — plus a long tail of framework-specific and tool-specific errors. Roughly seventy percent of the production incidents Afterbuild Labs audits trace back to four recurring failure classes: Supabase Row-Level Security policies missing or misconfigured on multi-tenant tables, Stripe webhook handlers that only process checkout.session.completed and ignore invoice.paid, customer.subscription.updated, and customer.subscription.deleted, OAuth redirect URLs still pointing at a preview subdomain after cutover, and environment variables split inconsistently between a local .env file and the Vercel dashboard. Each /fix page names the exact error string, explains the underlying cause in plain prose, gives the copy-paste code-level fix, and ships with a regression test so the same break does not reappear on the next deploy.