Northflank disclosed a security incident on October 30 after receiving an alert from cloud security company Wiz. The platform, which lets developers deploy applications, containers, databases and AI workloads without managing the underlying infrastructure, advised customers to immediately rotate their environment variables and API keys.
The company later told customers to assume that environment variables had been decrypted and that secrets stored within them may have been stolen. Such variables commonly contain database passwords, cloud credentials, authentication tokens and API keys. If exposed, they can be used to access other services, steal data or generate unauthorized charges.
Northflank subsequently confirmed that several customers had reported unauthorized use of AI API keys supplied to the platform. It advised users to rotate their stored secrets, review usage and billing logs from their AI providers, and report suspicious activity to both Northflank and those providers.
According to the company’s incident timeline, Wiz reported the breach at 11:43 UTC on October 30, and Northflank engineers reviewed the alert one minute later. The company applied a network policy patch and added further patches at 12:50 and 13:05 UTC. Northflank also requested help from Wiz’s incident response team.
Northflank has not disclosed how attackers accessed its systems, how many customers were affected, or whether credentials other than the reported AI API keys were stolen or abused. As part of its response, the company enabled credential rotation for PostgreSQL, MongoDB, MySQL and Redis. It also activated audit logs for all customers, including free-account users, to support investigations.
The company continues to recommend rotating secrets stored in workloads, secret groups and global secrets, as well as disabling public access to add-ons where that option is enabled.
The incident follows a series of breaches and exposures involving AI development platforms. In a report titled “Black October for AI Developer Platforms,” Wiz identified at least nine incidents during the month, including cases involving Railway, OpenCode, Warp and Lightning AI. Several incidents involved attempts to obtain AI service credentials stored in environment variables.
Wiz described the potential abuse as LLMJacking, in which stolen credentials are used to access AI models through another organization’s account. This can result in substantial usage charges or enable unauthorized resale of access. In two cases examined by the researchers, compromised enterprise credentials were used for approximately 2.5 million Amazon Bedrock API calls, producing around $100,000 in AI usage.
Wiz has not said that all nine incidents were carried out by the same attackers. Northflank said the unauthorized AI key activity reported by customers may be connected to the broader attacks identified by the security company.
Sources: BleepingComputer.