Apps built by Lovable's AI agent expose user data through missing database rules

In the wild Error Severity 3 significant Reported 74% EU Disputed

Lovable's AI app builder generated web apps whose browser code queried their Supabase databases directly with a public key, relying entirely on row-level security policies that in many apps were missing or insufficient. Replit employee Matt Palmer said he and a colleague scanned 1,645 Lovable-built apps showcased by Lovable and found 170 that let anyone read data such as users' names, email addresses, financial information and secret API keys; some also allowed unauthorised writes. Lovable responded with a security scan that, according to Palmer, checked only whether such policies existed, not whether they worked; the flaw was published as CVE-2025-48757.

Palmer, who found the problem in a Lovable-built app on 20 March 2025, flagged it to Lovable's CEO and said the company told him there was no issue; the CVE was made public after a 45-day disclosure window. Lovable said it was 'not yet where we want to be in terms of security'.

Disputed. Lovable disputes the CVE, saying each customer is responsible for protecting the data of their own application.

What would have stopped it

An AI app builder should generate server-side access checks by default and test that database rules actually block unauthorised reads before publishing.

Guardrails

1 of 8 failed
  1. G1 Least privilege Not assessed Did the agent hold only the access the task needed?
  2. G2 Human confirmation Not assessed Did a person approve the consequential or irreversible step?
  3. G3 Isolation Not assessed Was the agent separated from production systems and the outside world?
  4. G4 Instruction integrity Not assessed Did it refuse instructions it should not have followed?
  5. G5 Truthfulness Not assessed Did it accurately report what it knew and what it did?
  6. G6 Limits Not assessed Was there a cap on rate, spend, volume or scale?
  7. G7 Oversight Failed Exposed apps were found by outside researchers, not by Lovable or app owners.
  8. G8 Recovery Not assessed Could the damage be undone?

Sources

1 primary, 1 press
  1. CVE-2025-48757 Matt Palmer, 29 May 2025
    “This affects any Lovable project that uses a database, created on or before April 15th, 2025, but may also affect projects created after that date.”
    Primary Quote checked Archived copy
  2. “Of those, 170 allowed anyone to access information about the site’s users, including names, email addresses, financial information and secret API keys for AI services that would allow would-be hackers to run up charges billed to Lovable’s customers.”
    Press Quote checked Archived copy
  3. CVE-2025-48757 Tenable (CVE record), 30 May 2025
    “NOTE: this is disputed by the Supplier because each individual customer of the Lovable platform accepts a responsibility over protecting the data of their application.”
    Specialist Disputes Quote checked Archived copy

Confidence 74% comes from these sources alone. How we score

History

  1. 15 Oct 2026Entered in the register with 3 sources

Similar failures

All incidents
  1. EEAIR-2025-028 Oct 2025 Priceline's 'Penny' chatbot promises a refund on a non-refundable hotel and cancels it Priceline's Penny 73%
  2. EEAIR-2025-008 Apr 2025 Cursor's AI support agent 'Sam' invents a one-device login policy Cursor support agent 'Sam', Anysphere 97%
  3. EEAIR-2025-037 2025 Salt's Gomo chatbot wrongly says roaming data is included, leaving a customer with a bill Gomo customer-service AI chatbotSwitzerland 73%