Northflank Breach Leads to Unauthorized AI API Use

Northflank warned customers to rotate all stored secrets after reports of unauthorized use of AI API keys, according to updates on its status page. The incident was one of at least nine security incidents disclosed by AI developer platforms during October, according to research from Wiz.

Northflank instructed customers to assume all environment variables had been decrypted and to revoke and replace credentials stored in workloads, secret groups, and global secrets. Customers were also told to review AI provider usage and billing logs, report suspicious activity, and disable public access to add ons where enabled.

Wiz alerted Northflank at 11:43 UTC on October 30. Northflank reviewed the alert one minute later and deployed a network policy patch at 12:02 UTC, followed by further patches at 12:50 and 13:05. The company has since enabled credential rotation for PostgreSQL, MongoDB, MySQL, Redis, Minio and RabbitMQ add ons, and turned on audit logs for all customers, including free users.

Wiz lists Fireworks AI, Andromeda, Zeabur, Lightning AI, OpenCode, Railway, Warp, CommandCode and Northflank among the platforms that disclosed incidents in October. It says the attacks hunted for stored environment variables, with a focus on those granting AI access. Zeabur will shut down because of its incident, Wiz says. We reported earlier on the OpenCode and Railway incidents.

In two cases Wiz investigated, stolen enterprise keys were used for roughly 800,000 and 1.7 million Amazon Bedrock calls, which Wiz estimates likely ran up over $100,000 in usage at list price. Wiz linked the keys to services that resell discounted access to AI models, but has not established that the same attackers carried out all nine incidents. Northflank has not said how attackers got in or how many customers were affected, BleepingComputer reports.