Every table and policy checked
We list each table, view and function in exposed schemas and test policies as anonymous, authenticated and other-user roles. You see exactly which queries succeed that should not.
Supabase exposes your Postgres tables through an API that any browser can call with the public anon key. Whether your data is safe depends almost entirely on Row Level Security and a few settings that are easy to get wrong.
Project references from our team
+ 15,000 more
The anon key in your frontend is meant to be public. It is safe only because every table in an exposed schema should have Row Level Security enabled with policies that limit rows to the right user or role. If RLS is disabled on one table, or a policy simply returns true, anyone who opens the browser's developer tools can query that table directly.
This happens often in apps generated with Lovable, Bolt or similar tools, where tables are created quickly and policies are added later or not at all. Our AI Code Engineers at neooutput review Supabase projects systematically, write and test the missing policies, and leave you with a setup your team understands. For larger products, OnlineAppCreator can take over ongoing development and operations.
We list each table, view and function in exposed schemas and test policies as anonymous, authenticated and other-user roles. You see exactly which queries succeed that should not.
We check that the service role key and third-party secrets never reach the client and live only in edge functions or your backend. Exposed keys are rotated, not just removed.
Storage buckets need their own policies, and edge functions often skip JWT checks. Both are reviewed alongside the database.
Policies get automated tests in your repository, so a future migration or AI-generated change cannot silently reopen a hole.
You grant access to the project and repository. We work on a branch or staging project, never directly on production data.
We query the API with the public key and test accounts to find readable or writable data, then review policies, grants and security definer functions.
Policies, indexes and function changes are written as versioned migrations and reviewed with your team before deployment.
You receive a findings report with what was fixed, remaining risks and a checklist for future tables.
Agentic coding workshops and AI services. Hands-on agentic coding workshops for teams on Claude Code, OpenAI Codex and Cursor, remote or on-site, in English or German. Plus AI Code Engineers on demand who deploy, secure and scale AI-built apps.
Explore neooutput workshopsIdea to operating product, built by AI Code Engineers. We take ideas to operating digital products: discovery, AI-native engineering by our AI Code Engineers, launch, SLAs and scale. B2B and public sector, from an ISO 9001 and ISO 27001 certified team in Germany.
See how we build productsRelated service: Deploy, Secure and Scale
No. It is designed to be used in the browser and app. Security comes from Row Level Security and grants, which is why a missing policy is a serious problem.
Yes. Lovable and similar tools connect to a regular Supabase project, so we review it the same way. We also check the generated frontend for keys and calls that should run server side.
We test every policy change on staging first, using the same flows your users run. Changes go live only after your team has reviewed them, typically during a quiet period.
We look at region selection, data minimisation and access logging as part of the review. Legal assessment remains with your data protection officer, and we provide the technical facts they need.
15 minutes with an AI Code Engineer, free and without obligation. Or call our team around the clock.