FixVibe

// docs / baas security / supabase rls scanner

Supabase RLS scanner: check public access to your tables

A Supabase RLS scanner helps you review whether a deployed app exposes table data to signed-out visitors. Readable data is an observation, not proof that RLS is disabled: the database might have an intentional public policy. Use this guide to interpret a report, review grants and row-level policies together, and verify private-data boundaries with isolated test records.

What an external RLS scan can show

FixVibe's passive Supabase check looks for observable anonymous read access in a deployed app it can reach. It does not sign in as your users or use service-role privileges. Its results describe the resources and access observed during that scan, not every table or every possible identity.

Interpret the evidence in the context of your app's intended access rules:

  • Returned rows establish readable data. If a signed-out visitor receives records that should be private, investigate that exposure. The response alone does not tell you whether RLS is off, a policy permits the read, or another database object exposes the data.
  • HTTP 200 alone is not a data leak. A successful query can return an empty array because a policy filters the rows, a query matches nothing, or the table is empty. Check the returned records and their intended audience, not just the status code.
  • Read evidence does not establish write safety. A passive read-only scan does not verify INSERT, UPDATE or DELETE behavior, or prove isolation between signed-in users and tenants. Those boundaries need separate owner-controlled tests.
  • Public access can be intentional. A public product catalog may be meant for everyone; private orders are not. Treat exposed privileged credentials as a separate incident covered in the service-role-key guide.

Cosa non può dimostrare una scansione esterna

An outside-in result is not a complete policy audit. It cannot establish your intended audience, enumerate every access path, or prove every grant and policy correct. Review intentional public tables against the report. No finding can also reflect unavailable resources, incomplete discovery or untested access; it is not a security certificate.

Four policy review areas after a scan

Review these configuration risks in your own project. They are useful review areas, not a claim that an external scan can diagnose each one or that every AI coding tool creates them.

1. Exposed tables without RLS

For a table reachable through the Data API, grants decide which operations a role may attempt; RLS policies restrict which rows it may access. Missing RLS matters when the role also has access. Enable RLS on exposed tables and keep only the grants each client role needs. Review views and functions separately because table policies are not a universal boundary.

2. Public policies on private data

With RLS enabled, no applicable policy means default denial for roles subject to RLS. That is not itself a data leak. A broad read policy may be appropriate for public data but wrong for private records. Define which roles and rows should be readable before changing a policy to make an error disappear.

3. Missing user or tenant boundaries

For private user-owned or tenant-owned records, verify that policies compare access against the authenticated identity or trusted membership data. Review the whole expression and all applicable policies rather than copying a generic predicate. A policy that works for one user can still permit unintended cross-user access.

Shape 4: Overly permissive write policies

Review SELECT, INSERT, UPDATE and DELETE independently. A SELECT policy does not itself authorize INSERT. Restrict table grants and review both existing-row access and proposed new-row values where applicable. Exercise allowed and denied operations using disposable fixtures; do not test writes against customer records.

How to use a FixVibe RLS report

Use the report as one input to a broader review of the deployed app:

  1. Confirm the affected resource. Check that the reported project and readable tables belong to the deployment you intended to review.
  2. Decide whether the access is intended. Compare the observed anonymous access with your written public, user-owned and tenant-owned data rules.
  3. Review what remains untested. The passive RLS check does not create, update or delete database rows. Signed-in and write boundaries require separate verification.

The Supabase RLS finding identifies observed readable tables. Use that evidence to prioritize review without exporting sensitive records into tickets or AI prompts. Do not infer complete table coverage or a particular policy definition from a report.

Cosa fare quando lo scanner trova qualcosa

Prioritize confirmed exposure of private data promptly. Intentional public data needs an access decision, not an automatic shutdown. Make changes through reviewed migrations and verify expected application behavior:

  1. Define the expected access. Record which tables and fields are public, which are private, and which users or tenants may read or change each kind of record.
  2. Review grants and RLS together. Enable RLS on exposed tables, remove unnecessary privileges and add only the policies each role and operation requires. Keep privileged keys server-side. Use the official RLS guide rather than a universal copy-paste policy.
  3. Test both allow and deny cases. Use an isolated environment and disposable records for signed-out access, two separate users and, where relevant, two tenants. Check read, insert, update and delete expectations without touching customer data.
  4. Check the actual result, not only HTTP 200. A filtered empty result can be expected. An unauthorized identity receiving a private test record violates the intended boundary. Also confirm authorized users still receive the data and operations they should.
  5. Deploy, re-scan and retain policy tests. Re-scan the same deployment to check previously observed public access. A cleared finding does not replace signed-in, tenant or write tests, and a network failure is not proof of a fix.

Combine external scans with policy tests

An external scan observes reachable deployment behavior. Repository review examines policy and migration intent, while owner-run database tests exercise specific roles and records. These methods answer different questions. Use them together rather than treating a clean result from any one tool as proof that the database is secure.

Domande frequenti

Lo scanner leggerà o modificherà i miei dati?

The passive RLS check is read-only. It examines observable anonymous read access and reports limited evidence. It does not perform database write operations or prove that every user and tenant is correctly isolated.

Does no finding mean every Supabase table is secure?

No. Discovery, reachability and supported deployment configuration affect what a scan can observe. Paused or unavailable services and untested identities are not evidence of safe policies. Review grants, exposed database objects and role-specific tests even when no anonymous-read finding is reported.

Cosa succede se la mia chiave anon viene ruotata o la mia chiave pubblicabile cambia?

Changing a public key does not repair grants or policies. Review the access rules, deploy the correction and re-scan the current app. If a privileged key was exposed, follow the separate credential-containment process as well.

Does a public Supabase key mean my data is exposed?

No. Legacy anon keys and newer publishable keys are designed for public clients; users authenticate separately. Data access depends on the effective role, grants and policies. Do not assume every key format or deployment is covered by an external scan. A no-finding result does not prove all data is private.

Prossimi passi

Run a free FixVibe scan against your production URL — the Supabase RLS check runs on every plan including the free tier. For a deeper read on what else can leak from a Supabase project, see Supabase service role key exposed in JavaScript and Supabase storage bucket security checklist. For the umbrella view across all BaaS providers, read BaaS misconfiguration scanner.

Primary references

Use Supabase's RLS guide for grants, policies and role-based tests; the API-key guide for public and privileged key roles; and Supabase's empty-result explanation when a successful read returns no rows.

// scansiona la tua superficie baas

Trova la tabella aperta prima che lo faccia qualcun altro.

Inserisci una URL di produzione. FixVibe enumera i provider BaaS con cui parla la tua app, identifica i loro endpoint pubblici e riporta cosa un client non autenticato può leggere o scrivere. Gratis, senza installazione, senza carta.

  • Piano gratuito — 3 scansioni/mese, senza carta d'iscrizione.
  • Fingerprinting BaaS passivo — nessuna verifica del dominio necessaria.
  • Supabase, Firebase, Clerk, Auth0, Appwrite e altri.
  • Coding-agent prompts where code/config applies, plus provider-console steps for hosted BaaS fixes.
Esegui scansione BaaS gratuita →

nessuna registrazione richiesta

Supabase RLS scanner: check public access to your tables · FixVibe