We are releasing Agent Skills for Postgres Best Practices to help AI coding agents write high quality, correct Postgres code.

The Supabase Blog offers insights, tutorials, and updates on Supabase, an open-source alternative to Firebase for building scalable and real-time applications. Covering topics such as database management, authentication, and serverless functions, the blog provides resources for developers looking to leverage Supabase's features to build modern web and mobile apps. Developers can learn how to use Supabase to create secure, scalable, and performant applications that meet the needs of their users.

Supabase

Supabase released Agent Skills for Postgres Best Practices, a collection of 30 rules across 8 categories to help AI coding agents write correct, performant Postgres code. The rules follow the Agent Skills open standard and cover critical areas like query performance, connection management, Row Level Security, schema design, and concurrency. The repository addresses common mistakes seen across hundreds of thousands of Postgres projects, such as missing indexes, RLS bypasses, and connection pool exhaustion. These best practices complement the Supabase MCP server by teaching agents proper judgment while the MCP server handles database connections and execution.

Introducing: Postgres Best Practices

How this fits with the Supabase MCP server #

This are the missing parts of the AI puzzle:
<blockquote>
The agent treats the rules as authoritative guidance rather than generating advice from its training data.
</blockquote>
<blockquote>
The MCP server gives agents the ability to work with your database. These best practices teach them to do it correctly.
</blockquote>
Such things are really multiplying our productivity! Thanks @supabase!

Great move by Supabase.
Most AI agents can write SQL that works, but not SQL that’s safe, scalable, and production-ready.
Having Postgres best practices in an Agent Skills format is exactly what AI coding tools were missing.
This could prevent a lot of silent performance and RLS mistakes before they hit production.

As a newbie developer, I am overwhelming by technologies around the world, this is one of 100 posts that completely change me.

This is such a needed move 👏
Giving agents explicit Postgres judgment instead of just execution power is the real unlock. I love the framing — MCP as the steering wheel, and these best practices as the driving lessons. That’s exactly how agentic systems should be designed: capability + guardrails.
As someone building AI-powered apps on top of Supabase/Postgres, I’ve seen firsthand how easy it is for agents (and even devs) to miss RLS nuances, indexing trade-offs, or connection pool limits. Packaging production-grade DB wisdom into machine-readable skills is a very strong step toward reliable agent workflows.
Excited to experiment with this alongside MCP in my Next.js + Supabase stack — this is the direction AI-assisted backend engineering should go.