A sophisticated attack campaign dubbed “Greeter” has been discovered, revealing a new level of complexity in cloud-based attacks. Researchers at Era Security uncovered the campaign while investigating exposed cloud instances on Shodan, a search engine for internet-connected devices. What they found was a systematic and organized attack across multiple cloud providers, using AI infrastructure to assist in writing code and orchestrating attack steps.
The Greeter campaign began with an exploratory search for exposed cloud instances on Amazon DynamoDB, which revealed thousands of open instances vulnerable to a critical code-injection vulnerability. Era Security researchers then deployed AI-powered honeypots to catch attackers in action, mostly on Amazon Web Services (AWS). The results were astonishing – 90% of the nearly 3,000 scanned Flowise deployments were running unpatched versions vulnerable to the critical CVE-2024-58351.
The attackers’ methods are noteworthy. They mapped out privileges and the environment, hunted for credentials, exploited cloud metadata services across multiple providers to escalate access, and then locked in persistence. To avoid detection, they left seemingly benign “hello” messages in the logs, making it difficult for teams to find specific indicators of compromise. This repeatable, staged methodology is a departure from traditional opportunistic attacks.
The scale and intent behind Greeter are concerning. The campaign was deployed across several AWS regions at once, indicating a broader scope and potentially greater damage. Era Security warns that at least 50% of the 46 confirmed real-world victims remain connected to the Internet and visible, suggesting at least 20 organizations are currently exposed.
The sophistication levels of the threat actors also speak to the growing value of cloud and AI infrastructure as attack surfaces. The Greeter campaign is not an opportunistic smash-and-grab attack or a script kiddie – it’s a carefully planned operation that requires significant effort to launch. This highlights the need for defenders to shift their thinking about cloud security, incorporating a layered model built around network access control, data protection, and identity security.
The Greeter campaign also underscores how attackers are using AI infrastructure as a new way to establish persistence. A decade ago, cloud attacks targeted the Docker API, with threat actors exploiting workloads and escaping containers to reach the metadata server underneath. Now, they’re using AI-powered code to assist in writing and orchestrating attack steps.
The implications of Greeter for the wider industry are significant. It signals a shift towards easier-to-use and more sophisticated attack tools, making it harder for defenders to stay ahead. As threat actors continue to evolve, so must our defenses. This means incorporating AI-powered security measures and rethinking traditional approaches to cloud security.
For organizations using cloud services, this is a wake-up call – unpatched instances are vulnerable, and attackers will seize the opportunity. To mitigate this risk, ensure all cloud instances are up-to-date with the latest patches, implement robust network access control, and prioritize identity security. In an era where AI is increasingly being used for both legitimate and malicious purposes, it’s essential to stay one step ahead of the attackers.
Source: Dark Reading — 2026-11-11