Security scan
An AI review of your app's code and data permissions for exposed data, missing sign-in checks, unsafe writes and leaked keys, with one-press fixes.
Last updated September 17, 2026
On this page
The Security tab runs an AI review of your app before you publish and after big changes. It reads your latest code and every table's permissions and groups what it finds by how much it matters.
What it looks for
- Data that is more open than the app treats it: a table readable by anyone that the pages assume is private.
- Pages that hold personal data without a sign-in check.
- Writes a visitor could make that should need an account or an admin.
- Keys or secrets sitting in code.
Fixing
Each finding explains why it matters and how to fix it. Fix this issue and Fix all issues send the work to the builder. Ignore hides a finding you have decided is fine.
When to run it
Before your first publish, after any change to who can see what, and whenever the tab says the app has changed since the last scan. A scan uses a small amount of credits.
The scan is a review, not a guarantee. Permissions are the real lock, checked on every request. The scan makes sure they say what you think they say. See Data permissions.
Was this page helpful?