Nine AI Developer Platforms Hit in a Single Month, With Stolen AI Access Keys the Common Target

Wiz Research says at least nine AI developer platforms disclosed security incidents over the course of a single month. Fireworks, OpenCode and Warp have reported signs of an "agent swarm" attack, and Andromeda has described its incident as part of an "industry-wide attack." OpenCode and CommandCode have attributed their incidents to an attempt at "distillation." Zeabur has published a full write-up and says it will shut down because of the incident, while Railway has released a preliminary report and timeline.

Most incidents are still being analysed, but some details have emerged. Fireworks, Zeabur and Lightning all point to credential access as the root cause. Railway believes initial access came through privilege escalation in its build infrastructure. Warp suggests a third-party vendor dependency, and OpenCode points to a legacy system. All public reports describe targeted attacks that hunted stored environment variables, especially those granting access to AI services. No single vulnerability links the incidents.

Wiz says the picture fits recent reports of agentic attacks, which can help attackers compromise victims faster, and of LLMJacking, where stolen AI access is sold through a market of brokers. That market makes theft easier to profit from and may draw in more attackers. Reported impact so far is anecdotal, including an alleged theft of about $371,000 in cryptocurrency using keys stored in Railway. This article covers only the opening of the Wiz report, so further details may follow.