supabase-migration

name	supabase-migration
description	Write and apply Supabase database migrations safely. Use when creating tables, altering schema, adding RLS policies, or seeding data.
allowed-tools	Read, Write, Bash

Supabase Migration Skill

Procedure

1. Read existing migrations in supabase/migrations/ to understand current schema
2. Create new migration file: supabase/migrations/YYYYMMDDHHMMSS_description.sql
3. Write SQL with IF NOT EXISTS guards where appropriate
4. Always include RLS policies for customer-facing tables
5. Never delete or modify existing migration files — create new ones

RLS rules for this project

• customers: users can only read/update their own row
• orders: users can only read their own orders, insert with their own customer_id
• order_items: same as orders (join through order_id → customer_id check)
• fish_species, seasons, flight_windows, delivery_zones: public read, admin write
• inventory_availability: public read, admin write

Naming conventions

• Tables: snake_case, plural (fish_species, order_items)
• Columns: snake_case
• Foreign keys: referenced_table_id (e.g. fish_species_id, delivery_zone_id)
• Timestamps: created_at, updated_at (with default now())
• All monetary values: integer (AUD cents)

Gotchas

• Always check if table exists before ALTER
• Supabase AU region — never change project region
• RLS must be enabled AND policies created — enabling without policies blocks all access
• Seed data goes in supabase/seed.sql, not in migrations