Kazakhstan-linked Silent Lynx hackers used Hermes AI agents to steal 2,000+ Tajik government documents, CloudSek said
Singapore-based cybersecurity company CloudSek said a hacking group nicknamed Silent Lynx, which other researchers have tied to Kazakhstan, used Hermes, an open-source framework for autonomous AI agents, to steal more than 2,000 national security-related documents from Tajikistan. According to CloudSek, the hackers used Hermes for reconnaissance from around mid-August and broke into Tajikistan's electronic document system a couple of weeks later. The AI agent planned and ran the attacks and wrote a custom program to download the files and break them into chunks to steal them. CloudSek found the activity after the hackers left a server exposed to the internet. Kazakh and Tajik government officials did not immediately respond to requests for comment.
What the AI did
People in the Silent Lynx hacking group told Hermes, an open-source framework for autonomous AI agents, what to do instead of typing hacking commands by hand. According to CloudSek, Hermes planned the attacks, wrote the rules, ran them, kept notes and wrote up reports. After the hackers broke into Tajikistan's electronic document system, the AI agent wrote a custom program to download the files, organise them and break them into chunks to steal them.
First reported November 11, 2026 · Added to the register November 16, 2026 · 3 sources
- Deployer
- Silent Lynx
- Model
- Hermes
- Affected
- Tajikistan's electronic document system
- Country
- Tajikistan
- When it happened
- Not stated in the sources
- First reported
- November 11, 2026
What this means for you
Could this affect you?
If you run a government or electronic document system, particularly in Central Asia. Government bodies and organisations running electronic document systems, particularly in Central Asia, are exposed to spying groups that reportedly hand the scouting, break-in and data theft over to autonomous AI agents.
What to check
- Harden your document management systems against unauthorised access.
- Monitor document systems for automated scouting of weak points.
- Watch for bulk downloads or files being pulled out in small chunks.
- Add threat intelligence on AI agent attack frameworks such as Hermes to your detection rules.
Areas of your AI programme this touches
Timeline
- November 11, 2026First reported
- November 11, 2026CloudSek said it found evidence that the hackers spoke Russian, operated during Kazakh business hours and frequently accessed the exposed server through a residential internet connection in Kazakhstan.[1]
- November 12, 2026CloudSEK said the more than 2,000 stolen documents passed through Tajikistan's Defense Ministry and included flight clearance requests, records of military personnel transfers and reports on construction in border areas. According to the company's November 11 report, the hackers gained access to Tajikistan's electronic document system through a railway employee's account.[3]
Every fact and its source (6)
- Attribution evidenceRussian speakers, Kazakh business hours, Kazakh residential connection“worked Kazakh business hours and often accessed the exposed server through a residential Kazakh internet connection”[2]
- Researcher that reported the attackCloudSek (Singapore)“Singapore-based CloudSek said it caught the cyber spies”[2]
- Documents stolenMore than 2,000 national security-related documents“steal more than 2,000 national security-related documents”[2]
- How the activity was discoveredHackers left a server exposed to the internet“after they inadvertently left a server exposed to the internet”[2]
- Role of the AI agentPlanned and ran attacks, wrote rules, kept notes and reports“Hermes planned the attacks, wrote the rules, ran them, kept notes, and wrote up reports”[2]
- Types of documents stolenFlight clearance requests, military personnel transfers, border construction reports“The stolen documents included flight clearance requests, records of military personnel transfers, and reports on construction in border areas.”[3]
Sources
- Kazakhstan-linked hackers reportedly steal Tajik government documents using AInews.az · November 11, 2026
- Kazakhstan-linked hackers used AI to steal Tajik government documents, researchers sayeconomictimes.indiatimes.com · November 11, 2026
- Kazakhstan-Linked Hackers Used AI to Steal Tajik Defense Documents, Researchers Saytimesca.com · November 12, 2026
How this record is classified. Severity S (4/5): real harm to one organisation, data exposed, or a containment failure. OECD level: serious incident, an event in which an AI system led to serious harm. Evidence: Reported, meaning press reporting only.
The facts in At a glance and in Every fact and its source are quoted from the sources listed. The summary, What the AI did and What this means for you are written from those sources and checked against them automatically before publication. Records reflect what has been disclosed, not everything that has happened. Incidents are found through news feeds and the GDELT Project.

