The Invisible Threat: How AI Catches the Ursnif Trojan
20
Mar 2019
The cyber AI approach successfully detected the Ursnif infections even though the new variant of this malware was unknown to security vendors at the time.
Over the past few months, I’ve analyzed some of the world’s stealthiest trojan attacks like Emotet, which employ deception to bypass traditional security tools that rely on rules and signatures. Guest contributor Keith Siepel also explained how cyber AI defenses managed to catch a zero-day trojan on his firm’s network for which no such rules or signatures yet exist. Indeed, with the incidence of banking trojans having increased by 239% among our customer base last year, it appears that this kind of subterfuge is the new normal.
However, one particularly sophisticated trojan, Ursnif, takes deception a step further evidence of which we are still seeing emerge. Rather than writing executable files that contain malicious code, some of its variants instead exploit vulnerabilities inherent to a user’s own applications, essentially turning the victim’s computer against them. The result of this increasingly common technique is that — once the victim has been tricked into clicking a malicious link or duped into opening an attachment via a phishing email — Ursnif begins to ‘live off the land’, blending into the victim’s environment. And by exploiting Microsoft Office and Windows features, such as document macros, PsExec, and PowerShell scripts, Ursnif can execute commands directly from the computer’s RAM.
One of the most prevalent and destructive strains of the Gozi banking malware, Ursnif was recently placed at the center of a new campaign that saw it dramatically expand its functionality. Originally created to infect hosts with spyware in order to steal sensitive banking information and user credentials, it can now also deploy advanced ransomware like GandCrab. These new functions are aided by the elusive trojan’s aforementioned file-less capabilities, which render it invisible to many security tools and allow it to hide in plain sight within legitimate, albeit corrupted applications. Shining a light on Ursnif therefore requires AI tools that can learn to spot when these applications act abnormally:
Cyber AI detects Ursnif on multiple client networks
First campaign: February 4, 2019
Darktrace detected the initial Ursnif compromise on a customer’s network when it caught several devices connecting to a highly unusual endpoint and subsequently downloading masqueraded files, causing Darktrace’s “Anomalous File / Masqueraded File Transfer” model to breach. Such files are often masqueraded as other file types not only to bypass traditional security measures but also to deceive users — for instance, with the intention of tricking a user into executing a file received in a malicious email by disguising it as a document.
As it happens, this Ursnif variant was a zero-day at the time Darktrace detected it, meaning that its files were unknown to antivirus vendors. But while the never-before-seen files bypassed the customer’s endpoint tools, Darktrace AI leveraged its understanding of the unique ‘pattern of life’ for every user and device in the customer’s network to flag these file downloads as threatening anomalies — without relying on signatures.
A sample of the masqueraded files initially downloaded:
Following the initial suspicious downloads, the compromised devices were further observed making regular connections to multiple rare destinations not previously seen on the affected network in a pattern of beaconing connectivity. In some cases, Darktrace marked these external destinations as suspicious when it recognized the hostnames they queried as algorithm-generated domains. High volumes of DNS requests for such domains is a common characteristic of malware infections, which use this tactic to maintain communication with C2 servers in spite of domain black-listing. In other cases, the endpoints were deemed suspicious because of their use of self-signed SSL certificates, which cyber-criminals often use because they do not require verification by a trusted authority.
In fact, the large volume of anomalous connections commonly triggered a number of Darktrace’s behavioral models, including:
Compromise / DGA Beacon Anomalous Connection / Suspicious Self-Signed SSL Compromise / High Volume of Connections with Beacon Score Compromise / Beaconing Activity To Rare External Endpoint
Beaconing is a method of communication frequently seen when a compromised device attempts to relay information to its control infrastructure in order to receive further instructions. This behavior is characterized by persistent external connections to one or multiple endpoints, a pattern that was repeatedly observed for those devices that had previously downloaded malicious files from the endpoints later associated with the Ursnif campaign. While beaconing behavior to unusual destinations is not necessarily always indicative of infection, Darktrace AI concluded that, in combination with the suspicious file downloads, this type of activity represented a clear indication of compromise.
Figure 1: A device event log that shows the device had connected to internal mail servers shortly before downloading the malicious files.
Lateral movement and file-less capabilities
In the wake of the initial compromise, Darktrace AI also detected Ursnif’s lateral movement and file-less capabilities in real time. In the case of one infected device, an “Anomalous Connection / High Volume of New Service Control” model breach was triggered following the aforementioned suspicious activities. The device in question was flagged after making anomalous SMB connections to at least 47 other internal devices, and after accessing file shares which it had not previously connected. Subsequently, the device was observed writing to the other devices’ service control pipe – a channel used for the remote control of services. The anomalous use of these remote-control channels represent compelling examples of how Ursnif leverages its file-less capabilities to facilitate lateral movement.
Figure 2: Volume of SMB writes made to the service control pipe on internal devices by one of the infected devices, as shown on the Darktrace UI.
Although network administrators often use remote control channels for legitimate purposes, Darktrace AI considered this particular usage highly suspicious, particularly as both devices had previously breached a number of behavioral models as a result of infection.
Second campaign: March 18, 2019
A second Ursnif campaign was detected just this week. At the time of detection, no OSINT was available for the C2 servers nor the malware samples.
On a US manufacturer’s network, the initial malware download took place from: xqzuua1594[.]com/loq91/10x.php?l=mow1.jad hosted on IP 94.154.10[.]62. Every single malware download is unique. This is indicating auto-patching or a malware factory working in the background. Darktrace immediately identified this as another Anomalous File / Masqueraded File Transfer.
Directly after this, initial C2 was observed with the following parameters:
HTTP GET to: vwdlpknpsierra[.]email Destination IP: 162.248.225[.]14 URI: /images/CKicJCsNNNfaJwX6CJ/0Ohp3OUfj/pI_2FszUK7ybqh33Qdwz/bOUeatCG2Qfks5DTzzO/H6SeiL8YozEYXKfornjfVt/hBgfcPVPCOf1H/2qo12IGl/L3B18ld4ZSx37TbdTUpALih/A5dl8FVHel/jMPIKnQfd/H.avi User Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; Trident/7.0; rv:11.0) like Gecko
What’s interesting here is that the C2 server provides a Sufee Admin login page:
This C2 appears to have bad operational security (OPSEC) as browsing random URIs on the server reveals some of the dashboard’s contents:
The initial C2 communication was followed by sustained TCP beaconing to ksylviauudaren[.]band on 185.180.198[.]245 over port 443 with SSL encryption using a self-signed certificate. Darktrace highlighted this C2 behavior as Compromise / Sustained TCP Beaconing Activity To Rare Endpoint and Anomalous Connection / Repeated Rare External SSL Self-Signed IP.
As of the writing of this article, the domain ksylviauudaren[.]band was still not recognized in OSINT as malicious – highlighting again Darktrace’s independence of signatures and rules to catch previously unknown threats.
Conclusion
The cyber AI approach successfully detected the Ursnif infections even though the new variant of this malware was unknown to security vendors at the time. Moreover, it even managed to catch Ursnif’s file-less capabilities for lateral movement through its modelling of expected patterns of connectivity. In terms of the wider security context, the ease with which cyber AI flagged such sophisticated malware — malware which takes action by corrupting a computer’s own applications — further demonstrates that AI anomaly detection is the only way to navigate a threat landscape increasingly populated by near-invisible trojans.
Oops! Something went wrong while submitting the form.
Newsletter
Enjoying the blog?
Sign up to receive the latest news and insights from the Darktrace newsletter – delivered directly to your inbox
Thanks for signing up!
Look out for your first newsletter, coming soon.
Oops! Something went wrong while submitting the form.
Inside the SOC
Darktrace cyber analysts are world-class experts in threat intelligence, threat hunting and incident response, and provide 24/7 SOC support to thousands of Darktrace customers around the globe. Inside the SOC is exclusively authored by these experts, providing analysis of cyber incidents and threat trends, based on real-world experience in the field.
Author
Max Heinemeyer
Global Field CISO
Max is a cyber security expert with over a decade of experience in the field, specializing in a wide range of areas such as Penetration Testing, Red-Teaming, SIEM and SOC consulting and hunting Advanced Persistent Threat (APT) groups. At Darktrace, Max is closely involved with Darktrace’s strategic customers & prospects. He works with the R&D team at Darktrace, shaping research into new AI innovations and their various defensive and offensive applications. Max’s insights are regularly featured in international media outlets such as the BBC, Forbes and WIRED. Max holds an MSc from the University of Duisburg-Essen and a BSc from the Cooperative State University Stuttgart in International Business Information Systems.
Darktrace's Detection of State-Linked ShadowPad Malware
An integral part of cybersecurity is anomaly detection, which involves identifying unusual patterns or behaviors in network traffic that could indicate malicious activity, such as a cyber-based intrusion. However, attribution remains one of the ever present challenges in cybersecurity. Attribution involves the process of accurately identifying and tracing the source to a specific threat actor(s).
Given the complexity of digital networks and the sophistication of attackers who often use proxies or other methods to disguise their origin, pinpointing the exact source of a cyberattack is an arduous task. Threat actors can use proxy servers, botnets, sophisticated techniques, false flags, etc. Darktrace’s strategy is rooted in the belief that identifying behavioral anomalies is crucial for identifying both known and novel threat actor campaigns.
The ShadowPad cluster
Between July 2024 and November 2024, Darktrace observed a cluster of activity threads sharing notable similarities. The threads began with a malicious actor using compromised user credentials to log in to the target organization's Check Point Remote Access virtual private network (VPN) from an attacker-controlled, remote device named 'DESKTOP-O82ILGG'. In one case, the IP from which the initial login was carried out was observed to be the ExpressVPN IP address, 194.5.83[.]25. After logging in, the actor gained access to service account credentials, likely via exploitation of an information disclosure vulnerability affecting Check Point Security Gateway devices. Recent reporting suggests this could represent exploitation of CVE-2024-24919 [27,28]. The actor then used these compromised service account credentials to move laterally over RDP and SMB, with files related to the modular backdoor, ShadowPad, being delivered to the ‘C:\PerfLogs\’ directory of targeted internal systems. ShadowPad was seen communicating with its command-and-control (C2) infrastructure, 158.247.199[.]185 (dscriy.chtq[.]net), via both HTTPS traffic and DNS tunneling, with subdomains of the domain ‘cybaq.chtq[.]net’ being used in the compromised devices’ TXT DNS queries.
Figure 1: Darktrace’s Advanced Search data showing the VPN-connected device initiating RDP connections to a domain controller (DC). The device subsequently distributes likely ShadowPad-related payloads and makes DRSGetNCChanges requests to a second DC.
Figure 2: Event Log data showing a DC making DNS queries for subdomains of ‘cbaq.chtq[.]net’ to 158.247.199[.]185 after receiving SMB and RDP connections from the VPN-connected device, DESKTOP-O82ILGG.
Additional cases of ShadowPad were observed across Darktrace’s customer base in 2024. In some cases, common C2 infrastructure with the cluster discussed above was observed, with dscriy.chtq[.]net and cybaq.chtq[.]net both involved; however, no other common features were identified. These ShadowPad infections were observed between April and November 2024, with customers across multiple regions and sectors affected. Darktrace’s observations align with multiple other public reports that fit the timeframe of this campaign.
Darktrace has also observed other cases of ShadowPad without common infrastructure since September 2024, suggesting the use of this tool by additional threat actors.
The data theft thread
One of the Darktrace customers impacted by the ShadowPad cluster highlighted above was a European manufacturer. A distinct thread of activity occurred within this organization’s network several months after the ShadowPad intrusion, in October 2024.
The thread involved the internal distribution of highly masqueraded executable files via Sever Message Block (SMB) and WMI (Windows Management Instrumentation), the targeted collection of sensitive information from an internal server, and the exfiltration of collected information to a web of likely compromised sites. This observed thread of activity, therefore, consisted of three phrases: lateral movement, collection, and exfiltration.
The lateral movement phase began when an internal user device used an administrative credential to distribute files named ‘ProgramData\Oracle\java.log’ and 'ProgramData\Oracle\duxwfnfo' to the c$ share on another internal system.
Figure 3: Darktrace model alert highlighting an SMB write of a file named ‘ProgramData\Oracle\java.log’ to the c$ share on another device.
Over the next few days, Darktrace detected several other internal systems using administrative credentials to upload files with the following names to the c$ share on internal systems:
ProgramData\Adobe\ARM\webservices.dll
ProgramData\Adobe\ARM\wksprt.exe
ProgramData\Oracle\Java\wksprt.exe
ProgramData\Oracle\Java\webservices.dll
ProgramData\Microsoft\DRM\wksprt.exe
ProgramData\Microsoft\DRM\webservices.dll
ProgramData\Abletech\Client\webservices.dll
ProgramData\Abletech\Client\client.exe
ProgramData\Adobe\ARM\rzrmxrwfvp
ProgramData\3Dconnexion\3DxWare\3DxWare.exe
ProgramData\3Dconnexion\3DxWare\webservices.dll
ProgramData\IDMComp\UltraCompare\updater.exe
ProgramData\IDMComp\UltraCompare\webservices.dll
ProgramData\IDMComp\UltraCompare\imtrqjsaqmm
Figure 4: Cyber AI Analyst highlighting an SMB write of a file named ‘ProgramData\Adobe\ARM\webservices.dll’ to the c$ share on an internal system.
The threat actor appears to have abused the Microsoft RPC (MS-RPC) service, WMI, to execute distributed payloads, as evidenced by the ExecMethod requests to the IWbemServices RPC interface which immediately followed devices’ SMB uploads.
Figure 5: Cyber AI Analyst data highlighting a thread of activity starting with an SMB data upload followed by ExecMethod requests.
Several of the devices involved in these lateral movement activities, both on the source and destination side, were subsequently seen using administrative credentials to download tens of GBs of sensitive data over SMB from a specially selected server. The data gathering stage of the threat sequence indicates that the threat actor had a comprehensive understanding of the organization’s system architecture and had precise objectives for the information they sought to extract.
Immediately after collecting data from the targeted server, devices went on to exfiltrate stolen data to multiple sites. Several other likely compromised sites appear to have been used as general C2 infrastructure for this intrusion activity. The sites used by the threat actor for C2 and data exfiltration purport to be sites for companies offering a variety of service, ranging from consultancy to web design.
Figure 6: Screenshotof one of the likely compromised sites used in the intrusion.
At least 16 sites were identified as being likely data exfiltration or C2 sites used by this threat actor in their operation against this organization. The fact that the actor had such a wide web of compromised sites at their disposal suggests that they were well-resourced and highly prepared.
Figure 7: Darktrace model alert highlighting an internal device slowly exfiltrating data to the external endpoint, yasuconsulting[.]com.
Figure 8: Darktrace model alert highlighting an internal device downloading nearly 1 GB of data from an internal system just before uploading a similar volume of data to another suspicious endpoint, www.tunemmuhendislik[.]com
Cyber AI Analyst spotlight
Figure 9: Cyber AI Analyst identifying and piecing together the various steps of a ShadowPad intrusion.
Figure 10: Cyber AI Analyst Incident identifying and piecing together the various steps of the data theft activity.
As shown in the above figures, Cyber AI Analyst’s ability to thread together the different steps of these attack chains are worth highlighting.
In the ShadowPad attack chains, Cyber AI Analyst was able to identify SMB writes from the VPN subnet to the DC, and the C2 connections from the DC. It was also able to weave together this activity into a single thread representing the attacker’s progression.
Similarly, in the data exfiltration attack chain, Cyber AI Analyst identified and connected multiple types of lateral movement over SMB and WMI and external C2 communication to various external endpoints, linking them in a single, connected incident.
These Cyber AI Analyst actions enabled a quicker understanding of the threat actor sequence of events and, in some cases, faster containment.
Attribution puzzle
Publicly shared research into ShadowPad indicates that it is predominantly used as a backdoor in People’s Republic of China (PRC)-sponsored espionage operations [5][6][7][8][9][10]. Most publicly reported intrusions involving ShadowPad are attributed to the China-based threat actor, APT41 [11][12]. Furthermore, Google Threat Intelligence Group (GTIG) recently shared their assessment that ShadowPad usage is restricted to clusters associated with APT41 [13]. Interestingly, however, there have also been public reports of ShadowPad usage in unattributed intrusions [5].
The data theft activity that later occurred in the same Darktrace customer network as one of these ShadowPad compromises appeared to be the targeted collection and exfiltration of sensitive data. Such an objective indicates the activity may have been part of a state-sponsored operation. The tactics, techniques, and procedures (TTPs), artifacts, and C2 infrastructure observed in the data theft thread appear to resemble activity seen in previous Democratic People’s Republic of Korea (DPRK)-linked intrusion activities [15] [16] [17] [18] [19].
The distribution of payloads to the following directory locations appears to be a relatively common behavior in DPRK-sponsored intrusions.
Observed examples:
C:\ProgramData\Oracle\Java\
C:\ProgramData\Adobe\ARM\
C:\ProgramData\Microsoft\DRM\
C:\ProgramData\Abletech\Client\
C:\ProgramData\IDMComp\UltraCompare\
C:\ProgramData\3Dconnexion\3DxWare\
Additionally, the likely compromised websites observed in the data theft thread, along with some of the target URI patterns seen in the C2 communications to these sites, resemble those seen in previously reported DPRK-linked intrusion activities.
No clear evidence was found to link the ShadowPad compromise to the subsequent data theft activity that was observed on the network of the manufacturing customer. It should be noted, however, that no clear signs of initial access were found for the data theft thread – this could suggest the ShadowPad intrusion itself represents the initial point of entry that ultimately led to data exfiltration.
Motivation-wise, it seems plausible for the data theft thread to have been part of a DPRK-sponsored operation. DPRK is known to pursue targets that could potentially fulfil its national security goals and had been publicly reported as being active in months prior to this intrusion [21]. Furthermore, the timing of the data theft aligns with the ratification of the mutual defense treaty between DPRK and Russia and the subsequent accused activities [20].
Darktrace assesses with medium confidence that a nation-state, likely DPRK, was responsible, based on our investigation, the threat actor applied resources, patience, obfuscation, and evasiveness combined with external reporting, collaboration with the cyber community, assessing the attacker’s motivation and world geopolitical timeline, and undisclosed intelligence.
Conclusion
When state-linked cyber activity occurs within an organization’s environment, previously unseen C2 infrastructure and advanced evasion techniques will likely be used. State-linked cyber actors, through their resources and patience, are able to bypass most detection methods, leaving anomaly-based methods as a last line of defense.
Two threads of activity were observed within Darktrace’s customer base over the last year: The first operation involved the abuse of Check Point VPN credentials to log in remotely to organizations’ networks, followed by the distribution of ShadowPad to an internal domain controller. The second operation involved highly targeted data exfiltration from the network of one of the customers impacted by the previously mentioned ShadowPad activity.
Despite definitive attribution remaining unresolved, both the ShadowPad and data exfiltration activities were detected by Darktrace’s Self-Learning AI, with Cyber AI Analyst playing a significant role in identifying and piecing together the various steps of the intrusion activities.
Credit to Sam Lister (R&D Detection Analyst), Emma Foulger (Principal Cyber Analyst), Nathaniel Jones (VP), and the Darktrace Threat Research team.
Appendices
Darktrace / NETWORK model alerts
User / New Admin Credentials on Client
Anomalous Connection / Unusual Admin SMB Session
Compliance / SMB Drive Write
Device / Anomalous SMB Followed By Multiple Model Breaches
Survey findings: AI Cyber Threats are a Reality, the People are Acting Now
Artificial intelligence is changing the cybersecurity field as fast as any other, both on the offensive and defensive side. We surveyed over 1,500 cybersecurity professionals from around the world to uncover their attitudes, understanding, and priorities when it comes to AI cybersecurity in 2025. Our full report, unearthing some telling trends, is out now.
Nearly 74% of participants say AI-powered threats are a major challenge for their organization and 90% expect these threats to have a significant impact over the next one to two years, a slight increase from last year. These statistics highlight that AI is not just an emerging risk but a present and evolving one.
As attackers harness AI to automate and scale their operations, security teams must adapt just as quickly. Organizations that fail to prioritize AI-specific security measures risk falling behind, making proactive defense strategies more critical than ever.
Some of the most pressing AI-driven cyber threats include:
AI-powered social engineering: Attackers are leveraging AI to craft highly personalized and convincing phishing emails, making them harder to detect and more likely to bypass traditional defenses.
More advanced attacks at speed and scale: AI lowers the barrier for less skilled threat actors, allowing them to launch sophisticated attacks with minimal effort.
Attacks targeting AI systems: Cybercriminals are increasingly going after AI itself, compromising machine learning models, tampering with training data, and exploiting vulnerabilities in AI-driven applications and APIs.
Safe and secure use of AI
AI is having an effect on the cyber-threat landscape, but it also is starting to impact every aspect of a business – from marketing to HR to operations. The accessibility of AI tools for employees improves workflows, but also poses risks like data privacy violations, shadow AI, and violation of industry regulations.
How are security practitioners accommodating for this uptick in AI use across business?
Among survey participants 45% of security practitioners say they had already established a policy on the safe and secure use of AI and around 50% are in discussions to do so.
While almost all participants acknowledge that this is a topic that needs to be addressed, the gap between discussion and execution could underscore a need for greater insight, stronger leadership commitment, and adaptable security frameworks to keep pace with AI advancements in the workplace. The most popular actions taken are:
Implemented security controls to prevent unwanted exposure of corporate data when using AI technology (67%)
Implemented security controls to protect against other threats/risks associated with using AI technology (62%)
This year specifically, we see further action being taken with the implementation of security controls, training, and oversight.
For a more detailed breakdown that includes results based on industry and organizational size, download the full report here.
AI threats are rising, but security teams still face major challenges
78% of CISOs say AI-powered cyber-threats are already having a significant impact on their organization, a 5% increase from last year.
While cyber professionals feel more prepared for AI powered threats than they did 12 months ago, 45% still say their organization is not adequately prepared—down from 60% last year.
Despite this optimism, key challenges remain, including:
A shortage of personnel to manage tools and alerts
Gaps in knowledge and skills related to AI-driven countermeasures
Confidence in traditional security tools vs. new AI based tools
This year, 73% of survey participants expressed confidence in their security team’s proficiency in using AI within their tool stack, marking an increase from the previous year.
However, only 50% of participants have confidence in traditional cybersecurity tools to detect and block AI-powered threats. In contrast, 75% of participants are confident in AI-powered security solutions for detecting and blocking such threats and attacks.
As leading organizations continue to implement and optimize their use of AI, they are incorporating it into an increasing number of workflows. This growing familiarity with AI is likely to boost the confidence levels of practitioners even further.
The data indicates a clear trend towards greater reliance on AI-powered security solutions over traditional tools. As organizations become more adept at integrating AI into their operations, their confidence in these advanced technologies grows.
This shift underscores the importance of staying current with AI advancements and ensuring that security teams are well-trained in utilizing these tools effectively. The increasing confidence in AI-driven solutions reflects their potential to enhance cybersecurity measures and better protect against sophisticated threats.
The full report for Darktrace’s State of AI Cybersecurity is out now. Download the paper to dig deeper into these trends, and see how results differ by industry, region, organization size, and job title.
![Event Log data showing a DC making DNS queries for subdomains of ‘cbaq.chtq[.]net’ to 158.247.199[.]185 after receiving SMB and RDP connections from the VPN-connected device, DESKTOP-O82ILGG.](https://cdn.prod.website-files.com/626ff4d25aca2edf4325ff97/67d1dbda2020b168dc1d83c2_Screenshot%202025-03-12%20at%2012.07.41%E2%80%AFPM.png)
![Darktrace model alert highlighting an internal device slowly exfiltrating data to the external endpoint, yasuconsulting[.]com.](https://cdn.prod.website-files.com/626ff4d25aca2edf4325ff97/67d1dd09fe9cf54a67c48f13_Screenshot%202025-03-12%20at%2012.14.04%E2%80%AFPM.png)
![Darktrace model alert highlighting an internal device downloading nearly 1 GB of data from an internal system just before uploading a similar volume of data to another suspicious endpoint, www.tunemmuhendislik[.]com](https://cdn.prod.website-files.com/626ff4d25aca2edf4325ff97/67d1dd45f3743ebe06d81d89_Screenshot%202025-03-12%20at%2012.14.46%E2%80%AFPM.png)
