Threat actors are extensively using CAIRN, PentAGI, and other tools for attacks at machine-processing speeds
Hackers have begun extensively using autonomous multi-agent systems to automate cyberattacks, allowing them to breach corporate networks at significantly higher speeds. According to researchers, attackers are moving from manual prompt input to full-scale software pipelines capable of independently finding vulnerabilities and creating exploits. This was reported by Google Threat Intelligence Group citing Mandiant, according to the portal PromPolitInform.
Attackers are increasingly deploying specialized frameworks such as ARTEX, CAIRN, CyberStrikeAI, HexStrike AI, and PentAGI. These tools allow for the automation of the entire intrusion cycle: from target reconnaissance and vulnerability analysis to code release and execution without human intervention. Such solutions significantly accelerate operations, turning cyberattacks into a continuous and high-speed process.
Automating vulnerability discovery and exploitation
The use of autonomous systems allows hackers to quickly analyze patches and create exploits for gaps that are known but not yet patched in many systems. A particular threat is posed by the specialized “exploit_agent” toolkit, which independently compares software versions, identifies weaknesses, and automatically attacks servers. Such automation significantly reduces the time system administrators have to implement defenses after security updates are released.
Cyber espionage and financial fraud
Beyond technical attacks, artificial intelligence is being actively integrated into espionage and information operations. Researchers have documented instances where groups linked to the PRC have experimented with developing their own frameworks to automate espionage activities. Meanwhile, financially motivated criminals continue to utilize “LLMJacking”—the theft of credentials to access powerful AI models, which are then resold through underground services.
Defense strategies and response
While autonomous tools increase attack speed, they often sacrifice stealth, leaving behind specific digital footprints and generating non-typical network traffic. Corporations are taking measures to counter this threat by implementing special behavioral analysis rules and architectural frameworks to protect infrastructure. Experts emphasize that the development of such technologies requires organizations to transition to a level of operational response that matches machine-processing data speeds.
Illustrative photo: Madhav-Malhotra-003 / CC0