Supabase security check: RLS, storage and auth
Your Supabase anon key is meant to be public, so everything it can reach depends on your row-level security policies. ScanAnt finds the Supabase project behind your app, looks for a leaked service_role key, and, once you prove ownership, checks what an anonymous visitor can actually read or write.
Free, no signup. We only read what your site already shows every visitor.
What we check
Each item is tagged with the scan that reports it. The quick scan is free and needs no signup. The others run only on sites you have proved you own.
service_role key in public code
Quick scanThe admin key bypasses every row-level security rule. Finding it in a browser bundle is a critical issue. We decode only its role and never use it.
Project detection
Quick scanWe recognise your project URL and anon key. Both are normal to expose and are labelled that way.
Tables readable by anyone
Verified scanFor each table the API lists, one request for at most one or two rows shows whether the anon key can read it. We save table and column names and counts, not values.
Open schema and GraphQL
Verified scanWhether the API lists your tables to anyone and whether GraphQL introspection is on.
Storage buckets
Verified scanWhether anyone can list the files in a bucket, and which bucket names the public key can see.
Auth settings
Verified scanOpen sign-ups, sign-ups without e-mail or phone confirmation and anonymous sign-ins.
Functions, RPC, Realtime and writes
Deep scanEdge Functions your site calls, database functions callable without login or readable through GET, whether Realtime accepts an anonymous subscription, and an anonymous insert request the database should refuse.
SQL policies in your migrations
Code scanIn uploaded SQL we flag tables without RLS, policies that always allow, weak WITH CHECK, users who can change their own role or plan, storage policies without an owner check and SECURITY DEFINER functions.
- Quick scan: No signup. Reads only what every visitor receives.
- Verified scan: After you prove you own the site. Read-only.
- Deep scan: After you prove ownership and tick the deep option.
- Code scan: You upload your project as a ZIP.
No scan can promise to find every problem. A clean result means no known problems were found on what the scan could see.
Common mistakes
Row Level Security turned off on a table
With RLS off, the anon key can read and often change every row.
A policy with USING (true)
It looks like a rule but lets everyone through.
Any signed-in user can read every row
A policy that only checks the user is logged in, without comparing the owner, exposes all users' data to each other.
Users can edit their own role or plan
Without a restricted WITH CHECK or a separate table, a profile update becomes a privilege escalation.
A service_role key in the front end
It gives the browser full database access, ignoring all policies.
How to fix it
Enable RLS on every table in the public schema and add explicit policies that compare auth.uid() with the row owner.
Replace USING (true) policies and test as the anon and authenticated roles in the SQL editor.
If a service_role key was exposed, rotate it in the dashboard and move the call to a server-side function.
Make storage buckets private and write policies on storage.objects that check ownership.
Confirm e-mail verification, turn off anonymous sign-ins if unused, and rescan.
Questions
Is my anon key being public a problem?
No. That is how Supabase is designed. The problem appears only when your RLS rules are missing or too loose, so the key can read or write more than intended.
Does the scan read my users' data?
It makes at most one or two read requests per table as proof, saves only the fact of access, the table and column names and the row count, and never stores the values.
Can it change or delete my data?
No. Everything is read-only apart from one optional write probe, a request the database is expected to reject. If a probe could leave a trace, such as a test row, we tell you before it runs.
What about CVE-2025-48757?
It describes ineffective row-level security in many Lovable-built apps. The same RLS gaps are what our table checks look for in any Supabase project.
Can I scan before launch?
Yes. Upload your migrations in a ZIP for the code scan, which checks your SQL policies without touching the live database.
Related checks
Lovable security check: is your Lovable app safe to launch?
Supabase RLS, leaked keys and client-side trust in apps built with Lovable.
Open →Firebase security check: Firestore, Realtime Database and Storage
Open Firestore and Realtime Database, listable Storage and test-mode rules.
Open →Find API keys and secrets exposed on your website
Leaked Stripe, OpenAI, AWS and database keys in public JavaScript.
Open →Bolt.new security check for apps you built with Bolt
VITE_ secrets, open backend tables, source maps and headers in Bolt apps.
Open →
Run the scan now
Free for any public site, no signup. Need the verified, deep or code scan? Choose the level on the scan page.
Choose a scan level →