STUPID-2026-0031
Lovable-generated apps shipped without adequate Supabase row-level security, leaving databases readable and writable by anyone (CVE-2025-48757)
Why 9/10
About 170 live apps had databases open to anyone. The exposure was real; whether anyone other than the researcher took data is not established.
- Scope
- 4/4
- Many organisations or the public: a flaw in a widely used product, mass exposure, a large outage
- Recoverability
- 2/3
- Partly lost, or loss reported without saying whether it was recovered
- Did it happen
- 2/2
- Happened in real use
- Beyond the user
- 1/1
- Production systems, customers or third parties, or credentials and sensitive personal data
The score is the sum of these four factors. How each factor is scored · scored 2026-10-29
Instruction given
Build an app with authentication and per-user data access.
Expected behavior
Restrict each user to their own records; unauthenticated visitors should have no access to protected data.
Actual behavior
Apps generated by Lovable before 15 April 2025 could ship with insufficient Supabase row-level security (RLS) policies. Per the CVE, remote unauthenticated attackers could read or write arbitrary database tables of the generated sites.
Damage
A researcher's scan of apps from Lovable's public showcase found about 170 of 1,645 (roughly 10%) with tables readable or writable without authentication; reported exposed data included home addresses, personal debt amounts and API keys. Whether attackers copied data from these apps is not established. Lovable responded with a security scanner; reports note it checks that RLS exists, not that it is correct.
Full write-up
CVE-2025-48757 records that an insufficient database row-level security (RLS) policy in Lovable-generated sites let remote, unauthenticated attackers read or write arbitrary database tables. Lovable builds apps on Supabase, where RLS policies are what stop the public API key in the browser from reaching every row; apps generated before 15 April 2025 could ship without adequate policies.
A security researcher scanning apps from Lovable's public showcase found about 170 of 1,645 vulnerable, across roughly 300 endpoints, with exposed data reportedly including home addresses, personal debt amounts and API keys. Because the gap came from what the generator produced rather than from one developer's mistake, it repeated across every affected app. Lovable's response was a security scanner; secondary reports note it detects whether RLS exists rather than whether the policies are correct.
An earlier version of this record cited an unrelated article (about Moltbook) and described the flaw as inverted access-control logic; both were corrected on 2026-10-29 against the CVE record.
Classification
- Agent
- Lovable
- Failure mode
- Security Vulnerability
- Root cause
- Logic Error
- Domain
- Backend
- MIT AI Risk Repository (our mapping)
- 2.2 AI system security vulnerabilities and attacks · 2.1 Compromise of privacy by leaking or correctly inferring sensitive information taxonomy
- Source
- Security Research
Related incidents
Cite this record
StupidLLM. "STUPID-2026-0031: Lovable-generated apps shipped without adequate Supabase row-level security, leaving databases readable and writable by anyone (CVE-2025-48757)." AI Agent Incident Database, 2026. https://www.stupidllm.com/incident/STUPID-2026-0031/ (accessed 2026-10-29). CC BY 4.0.
← STUPID-2026-0030By IDSTUPID-2026-0032 →
Get told when an agent breaks something
We document AI agent failures daily, severity-scored against a published scale. Each Monday, if any scored 7 or above that week — deleted data, leaked secrets, broken production — you get one email with the sources. When nothing did, you get nothing.
This database is callable over MCP — query it from inside your agent.