AI Risk Wire

Cases › leaked private data › LR-2026-0838

Lovable AI-generated apps ship with unauthenticated Supabase databases exposing passport, payment and chat data

Preliminary Why: 1 report: Startup Fortune · subject to change · 1 independent report

Developing: a major or severe case captured in the last 72 hours. Later changes are logged on the corrections page.

Status: no action reported

Severity major
Reported 2 Nov 2026
Developer and model Lovable
Agent tool Lovable
Could do database write, code generation, deployment
Industry Software Publishers

No source states: loss or records, deployer.

Severity major: many third parties' personal data was exposed at unbounded scope, with 24,664 projects returning records without authentication and one forum leaking more than 200,000 chat messages and 80,000 user profiles.

Lovable, an AI coding platform whose agent builds apps, wires up a database and ships them live, generated thousands of apps whose Supabase backends returned records without authentication. The German research group Crissy Field, working with Handelsblatt, pulled 159,041 Lovable projects from Common Crawl covering apps built between December 2025 and September 2026 and queried their databases without credentials. Of 34,545 projects on Supabase, 24,664 returned records with zero authentication; email addresses were readable on 7,403 sites, one database exposed over 1,000 passport records, a forum leaked more than 200,000 chat messages and 80,000 profiles, and a CRM-style project exposed more than 280,000 customer emails. The findings point to an agent that, left to its defaults, generates working apps without enabling Row Level Security, despite Lovable's Security Scanner and AI-driven Security Reviewer added after CVE-2025-48757.

What happened

No source tells the story step by step yet.

What was done

No source states an action or fix.

Risks

Sorted by rule from the case's fields; not yet read from the sources.

Controls

Not yet read for controls. Controls that mitigate the primary risk:

Application

Data minimisation · Sensitive-data checks on outbound data · Human approval for high-impact actions · Output filtering and redaction · People told they are dealing with AI · Risk assessment per deployment

Infrastructure / target

Bot and scraping defences

Similar incidents

All cases of this kind ›

Evidence

No source states: a step-by-step account, when it happened, when it was found, the direct loss, any court case, what set it off, what was done.

Show every field, with its source and grade

The AI system

Developer
Lovable
C one secondary source states it
Tools it was given
database write, code generation, deployment
C one secondary source states it
Safeguards in place
automated Security Scanner, AI-driven Security Reviewer
C one secondary source states it
Stage of use
in live use
C one secondary source states it
How it was connected
agent with tools
C one secondary source states it
What it was for
describe an app in plain English and watch an AI agent build it, wire up a database and ship it live
C one secondary source states it
Meant for live use
yes
C one secondary source states it

What happened

Earliest source we hold
2026-11-02
C one secondary source states it
Who caught it
an outside party
C one secondary source states it
Themes
vibe coding, coding agent, data exposure, supabase rls, unauthenticated database, pii, passport data
C one secondary source states it
Cause
a safeguard that was in place did not hold
C one secondary source states it
Cause in detail
The AI agent generating backend code left Supabase Row Level Security unconfigured by default, and Lovable's Security Scanner and AI-driven Security Reviewer did not catch the misconfiguration before apps shipped.
C one secondary source states it
Who was responsible
the model provider
C one secondary source states it
Earlier warning signs
A separate flaw tracked as CVE-2025-48757 led the company to add an automated Security Scanner and an AI-driven Security Reviewer meant to catch exactly this kind of misconfiguration before it ships.
C one secondary source states it
What went wrong
leaked private data
C one secondary source states it
AI agent
an AI agent acted
C one secondary source states it
What it did
The agent generated and shipped apps whose databases were left in an insecure, publicly readable state without flagging it.
C one secondary source states it

Impact

Severity
major
C one secondary source states it
Kind of harm
privacy
C one secondary source states it
Who was affected
Thousands of vibe-coded apps reportedly exposed corporate and personal data on the open web.
C one secondary source states it
Kinds of data exposed
personal information, credentials
C one secondary source states it
Others exposed downstream
24,664
C one secondary source states it
Why this severity
many third parties' personal data was exposed at unbounded scope, with 24,664 projects returning records without authentication and one forum leaking more than 200,000 chat messages and 80,000 user profiles.
C one secondary source states it

Context

Industry
Software Publishers
C one secondary source states it
Base count
159,041 units, December 2025 to September 2026
C one secondary source states it
Shared exposure
one shared model or part put many users at risk
C one secondary source states it
Investigated
by an unpaid independent party
C one secondary source states it
How it was analysed
by people
C one secondary source states it
Insurance lines
cyber, privacy/liability
C one secondary source states it
Coverage
no precedent
C one secondary source states it
Notes
Loss falls on third-party data subjects whose records sat in unauthenticated databases generated by the platform's agent; allocation between platform provider and app builders is contested in the article.
C one secondary source states it

Notes on gaps

Extractor proposed values without quotes; dropped: .

LR-2026-0838 · version 1 · published 2026-11-04 after automatic checks; not yet checked field by field by a person · JSON