01

Your Supabase app can look ready.

The UI works. Login works. The demo feels fine. But one loose policy can still expose user data.

02

The risk is usually in the rules.

RLS, service keys, storage buckets, auth roles. These are easy to miss when you build fast.

03

So we made Database Sentinel.

A Claude Skill that audits the database and gives you the fix.

>_View on GitHub

Free · Open source · MIT licensed · made by Parth Jha (Farenhyte)

What it does

One audit pass for the database checks builders forget.

audit output

Find the backend

Detects Supabase, MongoDB, and the database files in the project.

database-sentinel audit

checks complete. report includes risk, proof, and fix.

From the lab

We built Database Sentinel while building Rehearsal AI on Supabase.

Rehearsal AI is a real product with real user data. As it grew on Supabase, one question kept coming back: do the database rules protect users as carefully as the interface suggests?

Database Sentinel came from that work. It checks the quiet parts of an app: RLS policies, storage rules, exposed keys, auth roles, and database access paths.

It sits beside Context Hub in the Jaipuria AI Labs open-source tooling layer around Rehearsal: one tool helps AI systems keep context, the other helps builders keep data boundaries.

Created byParth Jha·built withDr. Shiva Kakkarand the Rehearsal team at Gradeless AI.

FAQ

Things people ask before installing.

Database Sentinel is an open-source Claude Skill that checks whether your database is safe before launch. It audits policies, secrets, roles, storage rules, and access paths so fast-moving teams can catch data exposure before users do.

Audit the database before launch day.

Open source. MIT licensed. Built for teams shipping fast with AI, Supabase, MongoDB, and modern app stacks.