Warp has disclosed a breach involving a third-party infrastructure vendor that potentially exposed AI agent decision logs and user telemetry. Founder and CEO Zach Lloyd estimates that roughly 1% of Warp users may have been affected, with the investigation continuing.
According to BleepingComputer’s October 27 report, Lloyd said Warp became aware on Friday of what it believed was active, narrowly targeted exploitation of a vendor in its infrastructure. He said the company detected the operation within hours and took steps to remediate it and mitigate the impact. That account establishes when Warp discovered the activity, rather than when the attackers first gained access.
Warp combines a terminal, code editor and tools for directing and reviewing AI agents that write and test code and deploy software. The potentially affected information includes agent traces—records of agent activity and decisions—and telemetry. The company has not clarified whether the exposed data included source code, prompts, credentials or other sensitive information, leaving the precise consequences for affected developers unclear.
BleepingComputer reported that analysis of the incident suggested the attack used an “agent swarm” and what was described as “a very sophisticated penetration and escalation technique.” The report did not identify the vendor or explain the vulnerability or attack mechanism in technical detail.
Lloyd said Warp contacted the vendor about the vulnerability and notified users identified as affected. The company has also engaged CrowdStrike to assist with the investigation and informed law enforcement. The roughly 1% estimate remains provisional while that work continues.
