abdullah4ai/supabase

Supabase Integration

CRITICAL: Backend-First Execution Order

You MUST set up the Supabase backend BEFORE writing any Swift code. Follow this exact order:

Phase 1: Backend Setup (via Supabase MCP)

1. Create all tables with mcp__supabase__execute_sql — columns, types, foreign keys, constraints, indexes
2. Enable RLS on every table — ALTER TABLE ... ENABLE ROW LEVEL SECURITY
3. Create RLS policies for every table — SELECT, INSERT, UPDATE, DELETE as appropriate
4. Create storage buckets — public for images, private for documents
5. Create storage policies — per-user folder enforcement on storage.objects
6. Create triggers — for updated_at, denormalized counters, etc.
7. Verify with mcp__supabase__list_tables and mcp__supabase__list_storage_buckets
8. Check auth config with mcp__supabase__get_auth_config to confirm providers are enabled (auto-configured by nanowave)

Phase 2: Swift Code

8. Write Config/AppConfig.swift — Supabase URL + anon key
9. Write SupabaseService.swift — shared client
10. Write models with Codable + CodingKeys matching the table columns you just created
11. Write services that query the tables you just created
12. Write views

NEVER skip Phase 1. If you write Swift code that references tables that don't exist, the app will crash at runtime.

Client Initialization

Initialize SupabaseClient once in a shared service. Use PKCE auth flow (default for mobile).

import Supabase

@Observable
final class SupabaseService {
    static let shared = SupabaseService()
    let client: SupabaseClient

    private init() {
        client = SupabaseClient(
            supabaseURL: URL(string: AppConfig.supabaseURL)!,
            supabaseKey: AppConfig.supabaseAnonKey
        )
    }
}

AppConfig Pattern

Store Supabase credentials as static constants — injected by nanowave during build.

enum AppConfig {
    static let supabaseURL = "https://PROJECT_REF.supabase.co"
    static let supabaseAnonKey = "YOUR_ANON_KEY"
}

Key Rules

• Backend first — create tables, RLS, buckets via MCP before writing Swift code
• Never manage tokens manually — Supabase SDK auto-refreshes sessions
• Auth architecture is handled by the authentication skill (AuthService, guards, modes) — this skill covers only the Supabase auth API calls
• Data access architecture (repository protocols, DTOs, domain mapping) is handled by the repositories skill — this skill covers Supabase API patterns used inside concrete repository implementations
• Models use Codable (NOT @Model) — Supabase is the persistence layer, not SwiftData
• All operations are async/await — no callbacks, no Combine
• RLS on every table — never leave a table without Row Level Security
• Use XcodeGen MCP to add "Sign in with Apple" entitlement when auth is needed
• Auth providers are auto-configured by the nanowave pipeline — use mcp__supabase__get_auth_config to verify, use mcp__supabase__configure_auth_providers only if manual adjustment is needed

Available MCP Tools

• mcp__supabase__execute_sql — run SQL queries (SELECT, DML)
• mcp__supabase__list_tables — list tables in schemas
• mcp__supabase__apply_migration — track DDL as versioned migrations
• mcp__supabase__list_storage_buckets — list storage buckets
• mcp__supabase__get_project_url — get project URL for Swift client
• mcp__supabase__get_anon_key — get anon key for Swift client
• mcp__supabase__get_logs — query project logs
• mcp__supabase__configure_auth_providers — enable/disable auth providers (apple, google, email, phone, anonymous)
• mcp__supabase__get_auth_config — check current auth provider configuration
• mcp__supabase__set_secrets — set edge function environment variables (name/value pairs)
• mcp__supabase__list_secrets — list all project secrets
• mcp__supabase__delete_secrets — delete secrets by name

References

Core

• Schema Setup — table creation, types, foreign keys, triggers
• RLS Policies — Row Level Security patterns for every table type
• Auth Patterns — email auth, Apple Sign In, auth state, guards
• Database Patterns — CRUD, filtering, realtime subscriptions

Storage

• Storage Setup — bucket creation, storage policies, path conventions
• Storage Patterns — upload, download, public/signed URLs
• Storage Service — StorageService singleton, image compression, PhotosPicker flow, ViewModel upload patterns

Management API

• Secrets API — edge function environment variables (create, list, delete)
• Edge Functions — deploying and managing edge functions via API
• API Keys — retrieving project anon/service_role keys
• Realtime — enabling per-table realtime via SQL publication
• Webhooks & Triggers — database webhooks, pg_net, Vault integration