From August–September 2026, Chinese-speaking attackers used an open-source AI penetration testing framework called Cairn to autonomously conduct 25 operations against at least 19 organizations. The attackers weaponized Cairn by pairing it with Chinese AI models Zhipu GLM and DeepSeek.
The attackers supplied a target and a goal, then Cairn planned, executed and recorded results on its own. These attacks achieved unauthorized access or data retrieval at 10 of the targeted organizations, including:
- Accessing dozens of valid credentials
- Mass exposing personal data (exceeding one million records at one organization)
- Retrieving military and government identities
- Creating persistent administrator accounts in a production system
- Modifying a live industry standards-ballot document
Read more…
Source: Palo Alto Unit 42
Sign up for the Cyber Security Review Newsletter
The latest cyber security news and insights delivered right to your inbox
Related:
- Newly Registered Domains: Malicious Abuse by Bad Actors
August 20, 2019
Newly registered domains (NRDs) are known to be favored by threat actors to launch malicious campaigns. Academic and industry research reports have shown statistical proof that NRDs are risky, revealing malicious usage of NRDs including phishing, malware, and scam. Therefore, best security practice calls for blocking and/or closely monitoring NRDs in enterprise traffic. Despite the evidence, there hasn’t yet ...
- Post GandCrab, Cybercriminals Scouring the Dark Web for the Next Top Ransomware
August 19, 2019
A detailed look at underground forums shows that cybercriminals aren’t sure where to look on the heels of the GandCrab ransomware group shutting its doors – and low-level actors are taking advantage of that by developing their own strains. Ransomware continues to be a top threat, with Friday’s ransomware attack on 23 Texas local government and agencies and two ...
- Adwind Remote Access Trojan Hits Utilities Sector
August 19, 2019
Attackers are targeting entities from the utility industry with the Adwind Remote Access Trojan (RAT) malware via a malspam campaign that uses URL redirection to malicious payloads. Adwind (also known as jRAT, AlienSpy, JSocket, and Sockrat) is distributed by its developers to threat actors under a malware-as-a-service (MaaS) model and it is capable of evading detection by most major anti-malware ...
- Uncovering a MyKings Variant With Bootloader Persistence via Managed Detection and Response
August 19, 2019
In May, during the Managed Detection and Response service on-boarding process of an electronics company in the Asia-Pacific region, we noticed suspicious activity via the Trend Micro™ Deep Discovery™ Inspector that turned out to be related to EternalBlue, an exploit perhaps more popularly known for being used in the WannaCry attacks. After the discovery, we sent our first alert to the ...
- Hackers Use Fake NordVPN Website to Deliver Banking Trojan
August 19, 2019
The attackers who previously breached and abused the website of free multimedia editor VSDC to distribute the Win32.Bolik.2 banking Trojan have now switched their tactics. While previously they hacked legitimate websites to hijack download links infected with malware, the hackers are now creating website clones to deliver banking Trojans onto unsuspecting victims’ computers. This allows them to focus ...
- Router Network Isolation Broken By Covert Data Exfiltration
August 18, 2019
Software-based network isolation provided by routers is not as efficient as believed, as hackers can smuggle data between the networks for exfiltration. Most modern routers offer the possibility to split the network into multiple segments that work separately. One example is a guest network that works in parallel with the host. The boundary insulates sensitive or critical ...