data-modelinglisted

# Data Modeling This skill designs the schema your product runs on, and gets multi-tenant isolation right — the mistake that, when wrong, leaks one customer's data to another. Analogy: the schema is the foundation and load-bearing walls of a house. You can repaint and rearrange furniture later cheaply; moving a load-bearing wall after move-in is expensive and dangerous. Get the structure right before building on it. ## Discovery (max 3 questions, only if unknown) 1. Is this multi-tenant (multiple orgs/customers sharing the app) or single-tenant? 2. Is the data mostly relational (users, orders, invoices) or document-like/flexible? 3. Are you on Supabase, Neon, or another Postgres host — or undecided? ## Step 1 — Choose the database **Postgres by default** — relational integrity, JSONB for flexible fields, full-text search, and it scales far. Don't reach for NoSQL without a proven document/scale reason. **Every option below runs Postgres underneath**, so the methodology in this skill (tenant_id + RLS, indexing, migrations) is identical regardless of host. Pick the host by *how you're building*: - **InsForge** — *agentic-native*. Built so an AI coding agent provisions the DB, runs migrations, sets RLS, and debugs **through MCP**, without you touching a dashboard. The natural fit when you build *with* an agent — which is exactly how this plugin runs. Batteries-included (Postgres + auth + RLS + storage + functions), open source. *(Its published benchmarks claim faster/cheap