4 Sources
[1]
Kazakhstan-linked hackers used AI to steal Tajik government documents, researchers say
WASHINGTON, Nov 11 (Reuters) - A hacking group tied to Kazakhstan recently used AI to steal more than 2,000 national security-related documents from neighboring Tajikistan, a cybersecurity company said in a report it published on Wednesday, opens new tab. Singapore-based CloudSek said it caught
[2]
Kazakhstan-Linked Hackers Used AI to Steal Tajik Government Documents, Researchers Say
WASHINGTON, Nov 11 (Reuters) - A hacking group tied to Kazakhstan recently used AI to steal more than 2,000 national security-related documents from neighboring Tajikistan, a cybersecurity company said in a report it published on Wednesday. Singapore-based CloudSek said it caught the cyber
[3]
Kazakhstan-linked hackers used AI to steal Tajik government documents, researchers say
A hacking group tied to Kazakhstan recently used AI to steal more than 2,000 national security-related documents from neighboring Tajikistan, a cybersecurity company said in a report it published on Wednesday. Singapore-based CloudSek said it caught the cyber spies using Hermes - an open-source
[4]
Kazakhstan-linked hackers used AI to steal Tajik government documents, researchers say
WASHINGTON, Nov 11 (Reuters) - A hacking group tied to Kazakhstan recently used AI to steal more than 2,000 national security-related documents from neighboring Tajikistan, a cybersecurity company said in a report it published on Wednesday. Singapore-based CloudSek said it caught the cyber spies
Share
Copy Link
Kazakhstan-linked hackers used Hermes, an open-source AI framework, to autonomously steal more than 2,000 national security documents from Tajikistan. CloudSek researchers discovered the breach after the Silent Lynx group accidentally left a server exposed, revealing how AI cyber espionage is evolving beyond traditional cybercriminal operations.
A hacking group tied to Kazakhstan has deployed autonomous AI agents to steal more than 2,000 national security-related documents from neighboring Tajikistan, according to a report published by Singapore-based cybersecurity firm CloudSek
1
. The discovery marks a significant shift in AI cyber espionage, demonstrating that state-sponsored hackers are now outsourcing complex operations to autonomous hacking tools rather than relying solely on manual commands.CloudSek researchers uncovered the operation after the attackers inadvertently left a server exposed to the internet, providing a rare glimpse into how AI-driven cyberattacks are executed
2
. The exposed server revealed detailed instructions, operational logs, and the hackers' methodology for deploying AI in offensive operations against the Tajik government.The Kazakhstan hackers utilized Hermes, an open-source framework for autonomous AI agents, to orchestrate the entire attack lifecycle. "Instead of typing hacking commands by hand, the attacker told Hermes what to do, and Hermes planned the attacks, wrote the rules, ran them, kept notes, and wrote up reports," CloudSek told Reuters
3
. This approach represents a fundamental departure from traditional cyber operations, where human operators manually execute each step.Source: Market Screener
The AI agent didn't just follow pre-programmed scripts. After conducting reconnaissance beginning around mid-August, Hermes autonomously wrote a custom program to download files, organize them systematically, and break them into chunks for more effective data exfiltration
1
. The hackers successfully breached Tajikistan's electronic document system approximately two weeks after initiating reconnaissance activities.Among the exposed data were detailed instructions showing how the hackers configured their autonomous AI agent. The guidelines revealed a carefully crafted persona: "You are an elite offensive security specialist and red teamer," the instructions stated
4
. A red teamer is a cybersecurity professional who simulates covert attacks to test an organization's systems, suggesting the hackers wanted their AI to adopt sophisticated penetration testing methodologies.These instructions also included operational parameters for how the AI should communicate with its human operators, maintain documentation, and approach problem-solving during the breach. This level of customization indicates that AI-driven cyberattacks are becoming increasingly sophisticated, with attackers fine-tuning their tools for specific espionage objectives.
Related Stories
CloudSek attributed the operation to a group nicknamed Silent Lynx, which other researchers have previously tied to Kazakhstan
2
. The cybersecurity firm found multiple indicators pointing to Kazakh origins: the hackers spoke Russian, maintained work schedules aligned with Kazakh business hours, and frequently accessed the exposed server through a residential Kazakh internet connection1
.Neither Kazakh nor Tajik government officials responded to requests for comment, and efforts to reach Silent Lynx were unsuccessful
3
. The silence from both governments raises questions about diplomatic tensions and whether this incident represents an isolated operation or part of broader intelligence-gathering activities between the neighboring nations.This incident matters because it demonstrates that autonomous hacking tools are no longer limited to financially motivated cybercriminals. CloudSek researchers emphasized that their discovery shows "more advanced, espionage-focused hackers were also outsourcing their hacking sprees to AI"
1
. Reuters had previously reported on cybercriminals using autonomous AI agents, but this marks a documented case of state-linked actors adopting the same techniques for national security document theft.The use of an open-source AI framework like Hermes means these capabilities are accessible to a wide range of threat actors, not just well-funded intelligence agencies. Watch for increased adoption of AI-powered reconnaissance and custom program development in future breaches. Organizations should anticipate that attackers will continue automating complex attack chains, potentially accelerating the speed and scale of data exfiltration operations targeting national security documents and sensitive government systems.
Summarized by
Navi
24 Jul 2026•Technology

02 Sept 2026•Technology

12 Aug 2026•Technology
