Check your own app — free, in about a minute.
The free mini-scan checks your live homepage for TLS/HTTPS basics, response security headers, cookie flags, exposed sensitive files, Firebase/Supabase backend rules, and the WordPress plugin versions your site exposes. You get a score and the finding titles — no account, no card.
Run the free mini-scan →Firebase is the default backend for AI-built apps, and for good reason: describe the app, and the agent wires up a database, file storage, and sign-in before lunch. What the agent rarely does is slow down for the part that decides whether your users' data is public — the security rules.
Here is the uncomfortable fact about Firebase: your web config is not a secret. The apiKey, project ID, and database URL in your JavaScript are public by design — Google says so. The only thing standing between the internet and your data is your rules file. These are the five ways that file goes wrong in shipped-fast apps.
When a project starts in test mode, the Realtime Database rules are effectively:
"rules": { ".read": true, ".write": true }
Anyone on the internet can read and write your entire database with no account at all. Test mode is meant to expire or be replaced before launch; in AI-built apps it routinely survives to production because the app "works" — of course it works, for everyone, including people who aren't your users.
The fix: before launch, open the Firebase console → Realtime Database → Rules and read what is actually deployed. If a stranger with your database URL could read the tree in a browser, your rules — not your code — are the bug.
The next step up from test mode is requiring authentication:
"rules": { ".read": "auth != null", ".write": "auth != null" }
This feels secure and isn't. Sign-up is open in most apps, which means anyone can mint an account in seconds and then read everything — every user's records, not just their own. Authentication answers "who are you?" Authorization — "is this your data?" — is a separate check the rules have to make, usually by keying reads and writes to the user's UID.
The fix: scope rules per user (auth.uid === $uid on per-user paths) and deny by default everywhere else.
Realtime Database rules cascade: a .read granted at a parent path applies to everything beneath it. A permissive rule near the root — added once to make a dashboard or a list view work — quietly exposes the entire database tree, including subtrees the developer added months later and never re-checked.
The fix: grant reads at the deepest path that needs them, and re-audit rules whenever you add a new top-level collection. Rules you wrote for the data you had are still governing the data you have now.
Firestore or RTDB gets the careful rules; Firebase Storage gets allow read, write: if true so profile pictures and uploads "just work." Attackers don't care which door you left open. Public write access to Storage means strangers can host files on your project's bandwidth and read everything users uploaded.
The fix: apply the same per-user scoping to Storage rules as to the database, and check both in the console — they live in different tabs, which is exactly why one gets forgotten.
Rules that check who is writing but not what invite a second class of abuse: oversized payloads, wrong types, and fields your app never intended — including, in the classic pattern, a user editing their own record to set role: "admin" because the app reads privileges from a writable profile field.
The fix: validate shape and protect privilege-bearing fields. If the client can write it, assume a hostile client will.
None of these mistakes require your source code to find. Your Firebase config is in your own pages; from it, anyone can attempt the same unauthenticated shallow read an attacker would. That is what an outside-in check does: PreShipIt's scan finds the Firebase config in your HTML/JS, probes your own project for an open read, and records counts only — never your data. Open rules are a High finding; rules that deny the probe are recorded as a strength.
Related: [I ...[truncated 887 chars]
The PreShipIt Brief — a short weekly email for people shipping vibe-coded apps: the week's important vulnerabilities and launch-safety checks, in plain English.
New and free. Double opt-in: we email you a confirmation link before anything else arrives. Unsubscribe any time.
Check your own app — free, in about a minute.
The free mini-scan checks your live homepage for TLS/HTTPS basics, response security headers, cookie flags, exposed sensitive files, Firebase/Supabase backend rules, and the WordPress plugin versions your site exposes. You get a score and the finding titles — no account, no card.
Run the free mini-scan →← Back home · Run a free mini-scan · Scan my app — $29 · See a sample report