Cloud application hosting provider Northflank is urging customers to rotate every credential and secret stored on its platform after a security breach led to the unauthorized use of customer AI API keys. The company warned that environment variables may have been decrypted and their contents stolen, although it has not disclosed how the attacker gained access, how many customers were affected, or whether credentials other than AI API keys were misused.
Northflank enables developers to deploy and manage applications, containers, databases, and AI workloads without managing the underlying infrastructure.
In a security incident update published on October 30, the company said it warned customers to immediately rotate their environment variables and API keys after receiving a threat alert from cloud security company Wiz.
Northflank subsequently told customers to assume that all environment variables had been decrypted and that secrets stored in them had been stolen.
“Assume all environment variables are decrypted. Disable/rotate all keys,” the company warned in an October 30 update.
Environment variables commonly contain sensitive information, including database passwords, cloud service credentials, authentication tokens, and API keys. If exposed, those credentials could be used to access other customer services, steal data, or generate fraudulent charges through API usage.
Northflank later acknowledged that multiple customers had reported unauthorized use of AI API keys stored on the platform.
“Users have reported unauthorized use of AI API keys provided to the Northflank platform,” the company said.
“If you haven’t already rotated your user-provided secrets stored in Northflank, please do so now.”
The company advised customers to review usage and billing logs from their AI providers for unexpected activity. Customers were also asked to report suspicious usage to Northflank and the relevant AI provider.
Northflank applied network patches after Wiz alert
According to Northflank’s incident timeline, Wiz notified the company of the breach at 11:43 a.m. UTC on October 30. A Northflank engineer confirmed the incident one minute later.
Northflank responded with a network policy patch, followed by additional patches at 12:50 UTC and 13:05 UTC. The company also requested help from Wiz’s incident response team.
Northflank has enabled credential rotation for PostgreSQL, MongoDB, MySQL, and Redis. It also enabled audit logging for all customers, including those on free accounts, to support investigations into suspicious activity.
The company continues to advise customers to rotate secrets stored in workloads, secret groups, and global secrets. Customers should also disable public access to add-ons if that feature is enabled.
BleepingComputer contacted Northflank with questions about the breach, its impact on customers, and whether stored credentials were accessed or stolen, but the company had not responded.
Wiz reported multiple attacks on AI development platforms
The Northflank incident occurred during a series of attacks targeting AI development platforms. Security researchers believe these platforms are attractive targets because they store API keys that can provide access to costly AI services.
In a report titled “Black October, an AI developer platform,” Wiz identified at least nine security incidents affecting AI development platforms during the same month. The platforms named in the report included Northflank, Railway, OpenCode, Warp, and Lightning AI.
The Railway incident also created the potential for decrypted customer environment variables to be exposed, while OpenCode published leaked AI API keys. Warp reported a breach involving AI agent logs and user telemetry data.
According to Wiz, attackers in some of these incidents specifically targeted environment variables containing AI service credentials. Those credentials can be abused in a practice known as LLMjacking.
LLMjacking involves using stolen credentials to access AI models through another organization’s API account. Attackers can use the access to generate large bills or resell access to the compromised account.
In two cases investigated by Wiz, compromised enterprise credentials were used to make approximately 2.5 million Amazon Bedrock API calls, generating approximately $100,000 in AI usage charges.
Researchers also identified infrastructure linked to services that resell access to AI models. However, they could not connect those operations to the initial breaches or confirm that they were involved in large-scale exploitation.
Wiz did not confirm that the same attacker was responsible for all nine incidents, although many of the cases involved attempts to steal AI API keys.
Northflank referenced the Wiz investigation in its incident update and said customer reports of compromised AI keys could be related to the attacks identified by the researchers.
The overall prevention score can hide what happens after the initial access. If an attacker uses valid credentials, your defenses drop sharply.
The Blue Report 2026 measures defense techniques by technology across 338 million simulations run in customer production environments.
Source: www.bleepingcomputer.com