Malware on Google Play Targets North Korean Defectors

Earlier this year, McAfee researchers predicted in the McAfee Mobile Threat Report that we expect the number of targeted attacks on mobile devices to increase due to their ubiquitous growth combined with the sophisticated tactics used by malware author…

Earlier this year, McAfee researchers predicted in the McAfee Mobile Threat Report that we expect the number of targeted attacks on mobile devices to increase due to their ubiquitous growth combined with the sophisticated tactics used by malware authors. Last year we posted the first public blog about the Lazarus group operating in the mobile landscape. Our recent discovery of the campaign we have named RedDawn on Google Play just a few weeks after the release of our report proves that targeted attacks on mobile devices are here to stay.

RedDawn is the second campaign we have seen this year from the “Sun Team” hacking group. In January, the McAfee Mobile Research Team wrote about Android malware targeting North Korean defectors and journalists. McAfee researchers recently found new malware developed by the same actors that was uploaded on Google Play as “unreleased” versions. We notified both Google, which has removed the malware from Google Play, and the Korea Internet & Security Agency.

Our findings indicate that the Sun Team is still actively trying to implant spyware on Korean victims’ devices. (The number of North Korean defectors who came to South Korea exceeded 30,000 in 2016, according to Radio Free Asia.) Once the malware is installed, it copies sensitive information including personal photos, contacts, and SMS messages and sends them to the threat actors. We have seen no public reports of infections. We identified these malwares at an early stage; the number of infections is quite low compared with previous campaigns, about 100 infections from Google Play.

Malware on Google Play

Malware uploaded on Google Play (now deleted).

We found three apps uploaded by the actor we named Sun Team, based on email accounts and Android devices used in the previous attack. The first app in this attack, 음식궁합 (Food Ingredients Info), offers information about food; the other two apps, Fast AppLock and AppLockFree, are security related. 음식궁합 and Fast AppLock secretly steal device information and receive commands and additional executable (.dex) files from a cloud control server. We believe that these apps are multi-staged, with several components. AppLockFree is part of the reconnaissance stage we believe, setting the foundation for the next stage unlike the other two apps. The malwares were spread to friends, asking them to install the apps and offer feedback via a Facebook account with a fake profile promoted 음식궁합.

Links to Previous Operations

After infecting a device, the malware uses Dropbox and Yandex to upload data and issue commands, including additional plug-in dex files; this is a similar tactic to earlier Sun Team attacks. From these cloud storage sites, we found information logs from the same test Android devices that Sun Team used for the malware campaign we reported in January. The logs had a similar format and used the same abbreviations for fields as in other Sun Team logs. Further, the email addresses of the new malware’s developer are identical to the earlier email addresses associated with the Sun Team. The relationship among email addresses and test devices is explained in the following diagram.

The use of identical email addresses ties the two malware campaigns to the same attacker.

About the Actors

After tracking Sun Team’s operations, we were able to uncover different versions of their malware. Following diagram shows the timeline of the versions.

Timeline of different malware versions of Sun Team.

Timeline shows us that malwares became active in 2017. Sun Team’s only purpose is to extract information from devices as all of the malwares are spywares. Malwares on Google Play stayed online for about 2 months before being deleted.

In our post of the earlier attack by this actor, we observed that some of the Korean words found on the malware’s control server are not in South Korean vocabulary and that an exposed IP address points to North Korea. Also, Dropbox accounts were names from South Korean drama or celebrities.

In the new malware on Google Play, we again see that the Korean writing in the description is awkward. As in the previous operation, the Dropbox account name follows a similar pattern of using names of celebrities, such as Jack Black, who appeared on Korean TV. These features are strong evidence that the actors behind these campaigns are not native South Koreans but are familiar with the culture and language. These elements are suggestive though not a confirmation of the nationality of the actors behind these malware campaigns.

Sun Team’s test devices originate from various countries.

Moreover, we uncovered information about the attacker’s Android test devices and exploits they tried to use. The devices are manufactured in several countries and carry installed Korean apps, another clue that the threat actors can read Korean. The exploits codes were found uploaded on one of the cloud storages used by Sun Team which are modified versions of publicly available sandbox escape, privilege escalation, code execution exploits that added functions to drop their own Trojans on victims’ devices. The modified exploits suggest that the attackers are not skillful enough to find zero days and write their own exploits. However, it is likely just a matter of time before they start to exploit vulnerabilities.

Modified exploits installing the Sun Team’s Trojan.

The most concerning thing about this Sun Team operation is that they use photos uploaded on social network services and identities of South Koreans to create fake accounts. We have found evidence that some people have had their identities stolen; more could follow. They are using texting and calling services to generate virtual phone numbers so they can sign up for South Korean online services.


This malware campaign used Facebook to distribute links to malicious apps that were labeled as unreleased versions. From our analysis, we conclude that the actor behind both campaigns is Sun Team. Be cautious when installing unreleased or beta versions of any app. Also, check the number of downloads to see if an app is widely installed; avoid obscure apps.

McAfee Mobile Security detects this malware as Android/RedDawn.A, B. Always keep your mobile security application updated to the latest version.

The post Malware on Google Play Targets North Korean Defectors appeared first on McAfee Blogs.

Lazarus Resurfaces, Targets Global Banks and Bitcoin Users

This blog was written with support and contributions provided by Asheer Maholtra, Jessica Saavedra Morales, and Thomas Roccia.
McAfee Advanced Threat Research (ATR) analysts have discovered an aggressive Bitcoin-stealing phishing campaign by the intern…

This blog was written with support and contributions provided by Asheer Maholtra, Jessica Saavedra Morales, and Thomas Roccia.

McAfee Advanced Threat Research (ATR) analysts have discovered an aggressive Bitcoin-stealing phishing campaign by the international cybercrime group Lazarus that uses sophisticated malware with long-term impact.

This new campaign, dubbed HaoBao, resumes Lazarus’ previous phishing emails, posed as employee recruitment, but now targets Bitcoin users and global financial organizations. When victims open malicious documents attached to the emails, the malware scans for Bitcoin activity and then establishes an implant for long-term data-gathering.

HaoBao targets and never-before-seen implants signal to McAfee ATR an ambitious campaign by Lazarus to establish cryptocurrency cybercrime at a sophisticated level.


Beginning in 2017, the Lazarus group heavily targeted individuals with spear phishing emails impersonating job recruiters which contained malicious documents. The campaign lasted from April to October and used job descriptions relevant to target organizations, in both English and Korean language. The objective was to gain access to the target’s environment and obtain key military program insight or steal money. The 2017 campaign targets ranged from defense contractors to financial institutions, including crypto currency exchanges, however; much of this fake job recruitment activity ceased months later, with the last activity observed October 22, 2017.


On January 15th , McAfee ATR discovered a malicious document masquerading as a job recruitment for a Business Development Executive located in Hong Kong for a large multi-national bank. The document was distributed via a Dropbox account at the following URL:


This is the mark of a new campaign, though it utilizes techniques, tactics and procedures observed in 2017. This document had the last author ‘Windows User’ and was created January 16, 2018 with Korean language resources. Several additional malicious documents with the same author appeared between January 16 though January 24, 2018.

Document summary from Virus Total


Malicious job recruitment documents

Victims are persuaded to enable content through a notification claiming the document was created in an earlier version of Microsoft Word. The malicious documents then launch an implant on the victim’s system via a Visual Basic macro.

Malicious Microsoft Word document


Implants dropped in campaign

The document (7e70793c1ca82006775a0cac2bd75cc9ada37d7c) created January 24, 2018 drops and executes an implant compiled January 22, 2018 with the name lsm.exe (535f212b320df049ae8b8ebe0a4f93e3bd25ed79). The implant lsm.exe contacted which also resolves to dropped in campaign

The other malicious document ( a79488b114f57bd3d8a7fa29e7647e2281ce21f6) created January 19, 2018 drops the implant (afb2595ce1ecf0fdb9631752e32f0e32be3d51bb); which is 99% similar-to the lsm.exe implant.

This document was distributed from the following Dropbox URLs:

  • hxxps://
  • hxxps:// description.doc?dl=1
HTTP response for job description document

This implant (csrss.exe) compiled January 15, 2018 contacts an IP address which resolves to We identified that this domain was used to host a malicious document from a previous 2017 campaign targeting the Sikorsky program.

  • hxxp://

A third malicious document (dc06b737ce6ada23b4d179d81dc7d910a7dbfdde) created January 19, 2018 drops e8faa68daf62fbe2e10b3bac775cce5a3bb2999e which is compiled January 15, 2018. This implant communicates to a South Korean IP address which resolves to

McAfee ATR analysis finds the dropped implants have never been seen before in the wild and have not been used in previous Lazarus campaigns from 2017. Furthermore, this campaign deploys a one-time data gathering implant that relies upon downloading a second stage to gain persistence. The implants contain a hardcoded word “haobao” that is used as a switch when executing from the Visual Basic macro.

Malicious Document Analysis

The malicious document contains two payloads as encrypted string arrays embedded in Visual Basic macro code. The payloads are present as encrypted string arrays that are decrypted in memory, written to disk and launched in sequence (second stage malicious binary launched first and then the decoy document).

The VBA Macro code is self-executing and configured to execute when the OLE document (MS Word doc) is opened (via “Sub AutoOpen()”). The AutoOpen() function in the VBA Macro performs the following tasks in the sequence listed:

  • Decodes the target file path of the second stage binary payload. This file path is calculated based on the current user’s Temp folder location:


VB code to decrypt second stage filepath
  • Decodes the second stage binary in memory and writes it to the %temp%\.\lsm.exe file location
second stage binary (MZ) as an encrypted String Array in the VBA Macro
second stage binary (MZ) decoded in memory by the VBA Macro
  • After writing the second stage payload to disk the VBA code performs two important actions.
    • Runs the second stage payload using cmd.exe. This is done so that the cmd.exe process exists as soon as the payload is launched. This way a process enumeration tool cannot find the parent process => Smaller footprint.

cmdline for executing the second stage binary:

cmd.exe /c start /b <temp_dir_path>\.\lsm.exe /haobao

  • Adds persistence on the system by creating a shortcut in the user’s Startup folder with the correct cmdline arguments:

Link file command line: <temp_dir_path>\.\lsm.exe /haobao

Link File Name: GoogleUpdate.lnk

Trigger code for executing the second stage binary and establishing persistence


LNK file configuration for establishing persistence
  • Once the second stage payload has been launched, the VBA Macro proceeds to display a decoy document to the end user. This decoy document is also stored in the VBA Macro as an encrypted string array (similar to the second stage payload). The decoy document is again written to the user’s temp directory to the following filename/path:

<temp_dir_path>\.\Job Description.doc

Decoy Document decoded in memory by the VBA Macro
  • Once the decoy document has been written to disk, the VBA Macro sets its file attributes to System + Hidden
  • The decoy document is then opened by the malicious VBA Macro and the original malicious document’s caption is copied over to the decoy document to trick the end user into mistaking the decoy document for the original (malicious) document.
  • This activity, combined with the fact that the VBA Macro then closes the current (malicious) document, indicates that the VBA Macro aims to trick an unsuspecting user into thinking that the decoy document currently open is the original (malicious) document opened by the user.
  • Since the decoy document is a benign file and does not contain any macros the victim does not suspect any malicious behavior.

Implant Analysis

As part of the implant initialization activities the implant does the following;

  • Checks the string passed to it through command line
    • “/haobao” in case of 535f212b320df049ae8b8ebe0a4f93e3bd25ed79
    • “/pumpingcore” in case of e8faa68daf62fbe2e10b3bac775cce5a3bb2999e

If the malware does not find this string in its cmdline arguments, it simply quits without going any further.

  • Unwraps a DLL into memory and calls its one-and-only import using Reflective DLL injection. DLL information.

During our research, we discovered additional variants of the DLL file.

DLL information


  • As part of Reflective DLL loading the malware performs the following tasks on the DLL it has unwrapped in memory:
    • Copy the unwrapped DLL into new locations in its own memory space.
    • Build imports required by the DLL (based on the IAT of the DLL)
Imports builder code in malware for the DLL imports
  • Call the newly loaded DLL image’s Entry Point (DllMain) with DLL_PROCESS_ATTACH to complete successful loading of the DLL in the malware process.
DLL Entry Point Call from malware to finish loading of the DLL in memory
  • Call the actual malicious export in the DLL named “CoreDn”
Hardcoded DLL export name “CoreDn” in malware

All the malicious activities described below are performed by the DLL unless specified otherwise.

Data Reconnaissance

The implant has the capability of gathering data from the victim’s system. The following information will be gathered and sent to the command and control server.

  • Computer name and currently logged on user’s name, stored in the format

<ComputerName> \ <Username>

Malware obtaining the computer name and user name
  • List of all processes currently running on the system arranged in format

<Process Name>\r\n

<Process Name>\r\n

<Process Name>\r\n

<Process Name>\r\n

Malware collecting process information from endpoint
  • The presence of a specific registry key on the system


  • The malware appends an indicator (flag) specifying whether the above registry key was found in the user’s registry:

This key is checked again as part of the command and control communication and is sent as a duplicate value to the command and control in the HTTP POST request as well (explained in the below).

Malware checking for the presence of the registry key



In preparation of the exfiltration of information collected from the endpoint, the malware performs the following activities:

  • Encode the collected information using a simple byte based XOR operation using the byte key: 0x34.
  • Base64 encode (standard) the XORed data.
  • Again, check for the presence of the Registry Key: HKCU\Software\Bitcoin\Bitcoin-Qt


Command and Control Server Communication

Once the malware has performed all these activities it sends an HTTP POST request to the CnC server:

  • www[dot] for md5 BDAEDB14723C6C8A4688CC8FC1CFE668
  • www[dot] for md5 D4C93B85FFE88DDD552860B148831026


In the format:

HTTP POST to www[dot]




HTTP POST to www[dot]



idx= 20 (14h) if the Registry key does not exist; 24 (18h) if the key exists.

no= XORed + base64 encoded “<Computername> \ <username>”

mode= XORed + base64 encoded Process listing + Registry key flag

Command and control server domain


The persistence mechanism of the malware is performed only for the downloaded implant. Persistence is established for the implant via the visual basic macro code initially executed upon document loading by the victim. This persistence is also performed ONLY if the malware successfully executes the downloaded implant. The malware first tries to update the HKEY_LOCAL_MACHINE registry key.

If the update is unsuccessful then it also tries to update the HKEY_CURRENT_USER registry key. Value written to registry to achieve persistence on the endpoint:

Registry Subkey = Software\Microsoft\Windows\CurrentVersion\Run

Value Name = AdobeFlash

Value Content = “C:\DOCUME~1\<username>\LOCALS~1\Temp\OneDrive.exe” kLZXlyJelgqUpKzP

Registry based persistence of the second stage payload

Connections to 2017 campaigns

The techniques, tactics and procedures are very similar to the campaigns that targeted US Defense contractors, US Energy sector, financial organizations and crypto currency exchanges in 2017.

The same Windows User author appeared back in 2017 in two malicious documents 비트코인_지갑주소_및_거래번호.doc and 비트코인 거래내역.xls which were involved in crypto currency targeting. Furthermore, one of the implants communicates to an IP address that was involved in hosting malicious job description documents in 2017 involving the Sikorsky military program.

McAfee Advanced Threat research determines with confidence that Lazarus is the threat group behind this attack for the following reasons:

  • Contacts an IP address / domain that was used to host a malicious document from a Lazarus previous campaign in 2017
  • Same author appeared in these recent malicious documents that also appeared back in Lazarus 2017 campaigns
  • Uses the same malicious document structure and similar job recruitment ads as what we observed in past Lazarus campaigns
  • The techniques, tactics and procedures align with Lazarus group’s interest in crypto currency theft


In this latest discovery by McAfee ATR, despite a short pause in similar operations, the Lazarus group targets crypto currency and financial organizations. Furthermore, we have observed an increased usage of limited data gathering modules to quickly identify targets for further attacks. This campaign is tailored to identifying those who are running Bitcoin related software through specific system scans.


 Indicators of Compromise

MITRE ATT&CK techniques

  • Data encoding
  • Data encrypted
  • Command-Line Interface
  • Account discovery
  • Process Discovery
  • Query registry
  • Hidden files and directories
  • Custom cryptographic protocol
  • Registry Run Keys / Start Folder
  • Startup Items
  • Commonly used port
  • Exfiltration Over Command and Control Channel




  • hxxps://
  • hxxps:// description.doc?dl=1


  • dc06b737ce6ada23b4d179d81dc7d910a7dbfdde
  • a79488b114f57bd3d8a7fa29e7647e2281ce21f6
  • 7e70793c1ca82006775a0cac2bd75cc9ada37d7c
  • 535f212b320df049ae8b8ebe0a4f93e3bd25ed79
  • 1dd8eba55b16b90f7e8055edca6f4957efb3e1cd
  • afb2595ce1ecf0fdb9631752e32f0e32be3d51bb
  • e8faa68daf62fbe2e10b3bac775cce5a3bb2999e

McAfee Detection

  • BackDoor-FDRO!
  • Trojan-FPCQ!
  • RDN/Generic Downloader.x
  • RDN/Generic Dropper
  • RDN/Generic.dx

The post Lazarus Resurfaces, Targets Global Banks and Bitcoin Users appeared first on McAfee Blogs.

Latest Intelligence for March 2017

Number of blocked web attacks increases to highest level since July 2016 and Necurs botnet returns with new spam campaigns.続きを読む

Number of blocked web attacks increases to highest level since July 2016 and Necurs botnet returns with new spam campaigns.


Attackers target dozens of global banks with new malware

Watering hole attacks attempt to infect more than 100 organizations in 31 different countries.続きを読む

Watering hole attacks attempt to infect more than 100 organizations in 31 different countries.