A contemporary cyber campaign is using a deceptive method known as ClickFix to distribute a previously undocumented malware loader called DeepLoad, raising fresh concerns about newly engineered attack techniques.
Researchers from ReliaQuest report that the malware is designed with advanced evasion capabilities. It likely incorporates AI-assisted obfuscation to make analysis more difficult and relies on process injection to avoid detection by conventional security tools. Alarmingly, the malware begins stealing credentials almost immediately after execution, capturing passwords and active session data even if the initial infection stage is interrupted.
The attack chain starts with a ClickFix lure, where users are misled into copying and executing a PowerShell command via the Windows Run dialog. The instruction is presented as a solution to a problem that does not actually exist. Once executed, the command leverages “mshta.exe,” a legitimate Windows binary, to download and launch a heavily obfuscated PowerShell-based loader.
To conceal its true purpose, the loader’s code is filled with irrelevant and misleading variable assignments. This approach is believed to have been enhanced using artificial intelligence tools to generate complex obfuscation layers that can bypass static analysis systems.
DeepLoad is carefully engineered to blend into normal system behavior. It disguises its payload as “LockAppHost.exe,” a legitimate Windows process responsible for managing the system lock screen, making its activity less suspicious to both users and security tools.
The malware also attempts to erase traces of its execution. It disables PowerShell command history and avoids standard PowerShell functions. Instead, it directly calls underlying Windows system functions to execute processes and manipulate memory, effectively bypassing monitoring mechanisms that track PowerShell activity.
To further evade detection, DeepLoad dynamically creates a secondary malicious component. By using PowerShell’s Add-Type feature, it compiles C# code during runtime, generating a temporary Dynamic Link Library (DLL) file in the system’s Temp directory. Each time the malware runs, this DLL is created with a different name, making it difficult for security solutions to detect based on file signatures.
Another key technique used is asynchronous procedure call (APC) injection. This allows the malware to execute its payload within a legitimate Windows process without writing a fully decoded malicious file to disk. It achieves this by launching a trusted process in a suspended state, injecting malicious code into its memory, and then resuming execution.
DeepLoad’s primary objective is to steal user credentials. It extracts saved passwords from web browsers and deploys a malicious browser extension that intercepts login information as users type it into websites. This extension remains active across sessions unless it is manually removed.
The malware also includes a propagation mechanism. When it detects the connection of removable media such as USB drives, it copies malicious shortcut files onto the device. These files use deceptive names like “ChromeSetup.lnk,” “Firefox Installer.lnk,” and “AnyDesk.lnk” to appear legitimate and trick users into executing them.
Persistence is achieved through Windows Management Instrumentation (WMI). The malware sets up a mechanism that can reinfect a system even after it appears to have been cleaned, typically after a delay of several days. This technique also disrupts standard detection methods by breaking the usual parent-child process relationships that security tools rely on.
Overall, DeepLoad appears to be designed as a multi-functional threat capable of operating across several stages of a cyberattack lifecycle. Its ability to avoid writing clear artifacts to disk, mimic legitimate system processes, and spread across devices makes it particularly difficult to detect and contain.
The exact timeline of when DeepLoad began appearing in real-world attacks and the overall scale of its use remain unclear. However, researchers describe it as a relatively new threat, and its use of ClickFix suggests it could spread more widely in the near future. There are also indications that its infrastructure may resemble a shared or service-based model, although it has not been confirmed whether it is being offered as malware-as-a-service.
In a separate but related finding, researchers from G DATA have identified another malware loader called Kiss Loader. This threat is distributed through phishing emails containing Windows Internet Shortcut files. When opened, these files connect to a remote WebDAV server hosted on a TryCloudflare domain and download another shortcut that appears to be a PDF document.
When executed, the downloaded file triggers a chain of scripts. It starts with a Windows Script Host process that runs JavaScript, which then retrieves and executes a batch script. This script displays a decoy PDF to avoid suspicion, establishes persistence by adding itself to the system’s Startup folder, and downloads the Python-based Kiss Loader.
In its final stage, Kiss Loader decrypts and executes Venom RAT, a remote access trojan, using APC injection. The extent of this campaign is currently unknown, and it is not clear whether the malware is part of a broader malware-as-a-service offering. The threat actor behind the operation has claimed to be based in Malawi, although this has not been independently verified.
Cyber threats are taking new shapes every day. Attackers are increasingly combining social engineering, fileless execution techniques, and advanced obfuscation to bypass traditional defenses. This evolution highlights the growing need for continuous monitoring, stronger endpoint protection, and improved user awareness to defend against increasingly sophisticated attacks.
ClickFix attacks are rapidly becoming a favored tactic among advanced persistent threat (APT) groups from North Korea, Iran, and Russia, particularly in recent cyber-espionage operations. This technique involves malicious websites posing as legitimate software or document-sharing platforms. Targets are enticed through phishing emails or malicious advertising and then confronted with fake error messages claiming a failed document download or access issue.
Cybersecurity experts have uncovered a new version of KoiLoader, a malicious software used to deploy harmful programs and steal sensitive data. The latest version, identified by eSentire’s Threat Response Unit (TRU), is designed to bypass security measures and infect systems without detection.
How the Attack Begins
The infection starts with a phishing email carrying a ZIP file named `chase_statement_march.zip`. Inside the ZIP folder, there is a shortcut file (.lnk) that appears to be a harmless document. However, when opened, it secretly executes a command that downloads more harmful files onto the system. This trick exploits a known weakness in Windows, allowing the command to remain hidden when viewed in file properties.
The Role of PowerShell and Scripts
Once the user opens the fake document, it triggers a hidden PowerShell command, which downloads two JScript files named `g1siy9wuiiyxnk.js` and `i7z1x5npc.js`. These scripts work in the background to:
- Set up scheduled tasks to run automatically.
- Make the malware seem like a system-trusted process.
- Download additional harmful files from hacked websites.
The second script, `i7z1x5npc.js`, plays a crucial role in keeping the malware active on the system. It collects system information, creates a unique file path for persistence, and downloads PowerShell scripts from compromised websites. These scripts disable security features and load KoiLoader into memory without leaving traces.
How KoiLoader Avoids Detection
KoiLoader uses various techniques to stay hidden and avoid security tools. It first checks the system’s language settings and stops running if it detects Russian, Belarusian, or Kazakh. It also searches for signs that it is being analyzed, such as virtual machines, sandbox environments, or security research tools. If it detects these, it halts execution to avoid exposure.
To remain on the system, KoiLoader:
• Exploits a Windows feature to bypass security checks.
• Creates scheduled tasks that keep it running.
• Uses a unique identifier based on the computer’s hardware to prevent multiple infections on the same device.
Once KoiLoader is fully installed, it downloads and executes another script that installs KoiStealer. This malware is designed to steal:
1. Saved passwords
2. System credentials
3. Browser session cookies
4. Other sensitive data stored in applications
Command and Control Communication
KoiLoader connects to a remote server to receive instructions. It sends encrypted system information and waits for commands. The attacker can:
• Run remote commands on the infected system.
• Inject malicious programs into trusted processes.
• Shut down or restart the system.
• Load additional malware.
This latest KoiLoader variant showcases sophisticated attack techniques, combining phishing, hidden scripts, and advanced evasion methods. Users should be cautious of unexpected email attachments and keep their security software updated to prevent infection.
Researchers have found new advancements in the ViperSoftX info-stealing malware, which was first discovered in 2020. This malware has become more sophisticated, using advanced techniques to avoid detection. One of its new methods is using the Common Language Runtime (CLR) to run PowerShell commands within AutoIt scripts, which are spread through pirated eBooks. This clever approach helps the malware to hide within normal system activities, making it harder for security software to detect.
How ViperSoftX Spreads
ViperSoftX spreads through torrent sites by pretending to be eBooks. The infection starts when users download a RAR archive that includes a hidden folder, a deceptive shortcut file that looks like a harmless PDF or eBook, and a PowerShell script. The archive also contains AutoIt.exe and AutoIt script files disguised as simple JPG image files. When a user clicks the shortcut file, it sets off a series of commands, starting with listing the contents of “zz1Cover4.jpg.” These commands are hidden within blank spaces and executed by PowerShell, performing various malicious actions.
What the Malware Does
According to researchers from Trellix, the PowerShell code performs several tasks, such as unhiding the hidden folder, calculating the total size of all disk drives, and setting up Windows Task Scheduler to run AutoIt3.exe every five minutes after the user logs in. This ensures the malware remains active on infected systems. Additionally, the malware copies two files to the %APPDATA%MicrosoftWindows directory, renaming them to .au3 and AutoIt3.exe.
A sneaky aspect of ViperSoftX is its use of CLR to run PowerShell within AutoIt, a tool normally trusted by security software for automating Windows tasks. This allows the malware to avoid detection. ViperSoftX also uses heavy obfuscation, including Base64 encoding and AES encryption, to hide commands in the PowerShell scripts extracted from image decoy files. This makes it difficult for researchers and analysis tools to understand what the malware does.
Additionally, ViperSoftX tries to modify the Antimalware Scan Interface (AMSI) to bypass security checks. By using existing scripts, the malware developers can focus on improving their evasion tactics.
The malware's network activity shows it tries to blend its traffic with legitimate system activity. Researchers noticed it uses deceptive hostnames, like security-microsoft[.]com, to appear more trustworthy and trick victims into thinking the traffic is from Microsoft. Analysis of a Base64-encoded User-Agent string revealed detailed system information gathered from infected systems, such as disk volume serial numbers, computer names, usernames, operating system versions, antivirus product information, and cryptocurrency details.
Researchers warn that ViperSoftX is becoming more dangerous. Its ability to perform malicious actions while avoiding traditional security measures makes it a serious threat. As ViperSoftX continues to evolve, it's essential for users to stay alert and use strong security practices to protect their systems from such advanced threats.
In addition, Italian cybersecurity firm Cleafy researchers Federico Valentini and Alessandro Strino reported an ongoing financial fraud campaign since at least 2019 that leverages a new web-inject toolkit called drIBAN. The main goal of drIBAN fraud operations is to infect Windows workstations inside corporate environments, altering legitimate banking transfers performed by the victims and transferring money to an illegitimate bank account.
These accounts are either controlled by the threat actors or their affiliates, who are then tasked with laundering the stolen funds. The fraudulent transactions are often realized by means of a technique called Automated Transfer System (ATS) that's capable of bypassing anti-fraud systems put in place by banks and initiating unauthorized wire transfers from a victim's own computer.
The operators behind drIBAN have become more adept at avoiding detection and developing effective social engineering strategies, in addition to establishing a foothold for long periods in corporate bank networks. Furthermore, there are indications that the activity cluster overlaps with a 2018 campaign mounted by an actor tracked by Proofpoint as TA554 targeting users in Canada, Italy, and the U.K.
Organisations need to be aware of these threats and take immediate action to protect their systems from cyberattacks. The ACN has reported that dozens of Italian organisations have been likely affected by the global ransomware attack and many more have been warned to take action to avoid being locked out of their systems.
Mandiant Managed Defense has reliably resolved GOOTLOADER infections since January 2021. When spreading GOOTLOADER, malicious actors cast a wide net, affecting a variety of industrial verticals and geographical areas.
Gootkit Malware
The Gootkit Trojan is Javascript-based malware that carries out a number of malicious tasks, such as authorizing threat actors remote access, recording video, capturing keystrokes, stealing emails, stealing passwords, and having the ability to inject malicious files to steal online banking login details.
Gootkit previously spread malware in the disguise of freeware installers, but now it deceives users into downloading these files by presenting them as legal documents. A user enters a search query into a search engine to begin the attack chain.
Mandiant Managed Defense believes that UNC2565, a group it tracks, is the sole group that the GOOTLOADER virus and infrastructure belong to at this time. Due to these breaches' rapid detection and mitigation, Mandiant's observation of post-compromise GOOTLOADER activities has mostly been restricted to internal surveillance.
If the GOOTLOADER file is successfully executed, other payloads like FONELAUNCH and Cobalt Strike BEACON or SNOWCONE that are saved in the registry will be downloaded. Future phases include PowerShell being used to execute these payloads.
The. NET-based loader FONELAUNCH is intended to load an encoded payload into memory, while the downloader SNOWCONE is responsible for obtaining next-stage payloads, notably IcedID, through HTTP.
The primary aims of Gootkit have remained the same, however, the attack process has undergone substantial modifications. Currently, the JavaScript file contained in the ZIP archive is trojanized and contains a different JavaScript file that is obfuscated and then begins to execute the malware.
Furthermore, to avoid detection, the malware's creators allegedly used three distinct strategies to cloak Gootkit, including hiding the code inside modified versions of trustworthy JavaScript libraries like jQuery, Chroma.js, and Underscore.js. These modifications show how actively developing and expanding UNC2565's capabilities remain.