AI agent flagged SQLite zero-day before in-the-wild exploitation

AI agent flagged SQLite zero-day before in-the-wild exploitation

First time an AI agent was credited with pre-empting active in-the-wild exploitation of a vulnerability.

July 2025HappenedCyber operations
Big Sleep (Google DeepMind / Project Zero AI agent) · Google · Global
What happened
In July 2025 Google reported that its Big Sleep AI agent discovered CVE-2025-6965, a critical memory-corruption vulnerability in SQLite (CVSS 7.2, affecting versions before 3.50.2) that Google said was known only to threat actors. Combined with Google threat intelligence indicating imminent exploitation, Google said it was able to cut off the vulnerability beforehand, describing it as the first time an AI agent was used to directly foil an attempt to exploit a vulnerability in the wild. Big Sleep had found its first real-world vulnerability around November 2024.
Response
SQLite was patched; Google expanded Big Sleep's use and later reported additional findings.
Discovered by
Big Sleep agent combined with Google Threat Intelligence.
Who was affected
SQLite (used widely); potential victims of imminent exploitation
Patterns
Autonomous cyber operation
Confidence: high · checked 24 September 2026. Published under CC BY 4.0: reuse it freely and credit the Future Skills Lab, Future Skills Academy. Seen an error or a newer development? Write to tellmemore@futureskills.academy.

Related

Incidents with the same pattern

What does this ask of people?

Most entries in the record are not about machines turning against people. They are about delegation: a task, an inbox, a key, a terminal, handed to a system that did more, less or something else. Futures literacy starts with noticing that.

All 177 incidents →The Future Skills Lab

Learn with us

Learn the human skills that gain value

Future Skills 101: live online sessions of 90 minutes, open to anyone. Power Skills trainings for teams on request.