Supabase

Supabase Plugin

Comprehensive Supabase integration for AI applications with complete database, auth, storage, realtime, and vector search capabilities.

Overview

This plugin provides high-level orchestration for Supabase using the cloud-hosted MCP server at https://mcp.supabase.com/mcp. It generates AI-optimized database schemas, configures authentication and security, sets up vector search with pgvector, and enables realtime features.

Features

• AI-First Database Schemas - Optimized for chat, RAG, agents, embeddings
• Vector Search - pgvector setup with hybrid search (semantic + keyword)
• Authentication - All OAuth providers, email auth, magic links
• Row Level Security - Template-based RLS policies for multi-tenant apps
• Realtime - Subscriptions, presence, broadcast
• Storage - File upload with CDN and policies
• Edge Functions - AI model integration (embeddings, LLMs)
• Supabase UI - Pre-built React components
• Type Generation - Auto-generate TypeScript types
• Comprehensive Testing - E2E workflows, schema validation, RLS testing

Installation

This plugin requires the Supabase MCP server to be configured in your .mcp.json:

{
  "mcpServers": {
    "supabase": {
      "type": "http"
      "url": "https://mcp.supabase.com/mcp?project_ref=${SUPABASE_PROJECT_REF}"
      "headers": {
        "Authorization": "Bearer ${SUPABASE_ACCESS_TOKEN}"
      }
    }
  }
}

Set environment variables:

• SUPABASE_PROJECT_REF - Your Supabase project reference ID
• SUPABASE_ACCESS_TOKEN - Your Supabase access token

Quick Start

Initialize Supabase in Project

/supabase:init

This command will:

• Configure Supabase in your project
• Set up environment variables
• Initialize Supabase CLI
• Create initial configuration

Create Complete AI App

/supabase:init-ai-app

This creates a complete AI application setup including:

• User management with auth
• Conversation/chat schema
• pgvector for embeddings
• RLS policies
• Realtime subscriptions

Commands

Initialization

• /supabase:init - Initialize Supabase in project
• /supabase:init-ai-app - Complete AI app setup
• /supabase:init-chat-app - Chat-specific setup
• /supabase:init-rag-app - RAG-specific setup

Features

• /supabase:add-auth - Add authentication
• /supabase:add-storage - Configure storage buckets
• /supabase:add-realtime - Setup realtime subscriptions
• /supabase:setup-pgvector - Configure vector search
• /supabase:setup-ai - Complete AI features setup
• /supabase:add-ui-components - Install Supabase UI

Schema & Database

• /supabase:create-schema - Generate database schema
• /supabase:add-rls - Add Row Level Security
• /supabase:deploy-migration - Deploy migrations
• /supabase:generate-types - Generate TypeScript types

Testing & Validation

• /supabase:validate-setup - Validate configuration
• /supabase:test-e2e - Run end-to-end tests
• /supabase:validate-schema - Validate database schema
• /supabase:test-rls - Test RLS policies
• /supabase:test-realtime - Test realtime features

Architecture

Multi-Layer Agent System

Execution Agents (MCP Wrappers)

• database-executor - Direct SQL execution
• migration-applier - Migration deployment
• project-manager - Project-level operations

Review Agents (Validation)

• code-reviewer - SQL/migration review
• security-auditor - RLS policy validation
• performance-analyzer - Query optimization

Expert Agents (Orchestration)

• architect - Schema design for AI apps
• security-specialist - Auth & RLS implementation
• realtime-builder - Realtime subscriptions
• ai-specialist - pgvector, embeddings, AI edge functions
• ui-generator - Supabase UI integration

Testing Agents

• validator - Setup validation
• tester - End-to-end workflows
• schema-validator - Database integrity

Skills

Each skill provides scripts, templates, and examples that agents use via Bash/Read tools: 1. pgvector-setup - Vector search configuration

• Scripts: setup-pgvector.sh, create-indexes.sh, setup-hybrid-search.sh, test-vector-search.sh
• Templates: Embedding tables, HNSW/IVFFlat indexes, hybrid search queries
• Use: Setting up vector search, RAG systems, semantic search

1. rls-templates - Row Level Security patterns

   • Scripts: generate-policy.sh, apply-rls-policies.sh, test-rls-policies.sh, audit-rls.sh
   • Templates: User isolation, multi-tenant, role-based access, AI chat policies
   • Use: Implementing RLS, securing multi-tenant apps

2. auth-configs - Authentication provider configs

   • Scripts: setup-oauth-provider.sh, setup-email-auth.sh, configure-jwt.sh, test-auth-flow.sh
   • Templates: OAuth configs (Google, GitHub, Discord), email templates, auth middleware
   • Use: Setting up authentication, OAuth providers, MFA

3. schema-patterns - AI application schemas

   • Scripts: generate-schema.sh, apply-migration.sh, validate-schema.sh, seed-data.sh
   • Templates: Chat schema, RAG schema, multi-tenant schema, AI usage tracking
   • Use: Designing database schemas for AI apps

4. e2e-test-scenarios - Complete test workflows

   • Scripts: run-e2e-tests.sh, test-auth-workflow.sh, test-ai-features.sh, test-realtime-workflow.sh
   • Templates: Test suites, CI/CD configs, test cases
   • Use: E2E testing, pre-deployment validation

5. schema-validation - SQL validation tools

   • Scripts: validate-sql-syntax.sh, validate-naming.sh, validate-constraints.sh, validate-indexes.sh, validate-rls.sh
   • Templates: Validation rules, naming conventions
   • Use: Validating schemas before deployment

6. rls-test-patterns - RLS testing patterns

   • Scripts: test-user-isolation.sh, test-multi-tenant-isolation.sh, test-role-permissions.sh, audit-rls-coverage.sh
   • Templates: Security test plans, vulnerability checklists
   • Use: Security testing, RLS validation

Documentation

550+ pages of Supabase documentation automatically fetched by agents via WebFetch.

Key documentation areas:

• Database: Tables, migrations, functions, triggers
• Auth: OAuth providers, RLS, sessions
• AI: pgvector, embeddings, semantic search
• Realtime: Subscriptions, presence, broadcast
• Storage: Buckets, policies, CDN
• Edge Functions: AI models, Deno runtime

Languages Supported

• TypeScript (Next.js, React, Node.js)
• Python (FastAPI, Django)
• JavaScript (vanilla, Node.js)

Version

1.0.0

License

MIT