rootxhacker/mistralai-7B-attack2ttp
Updated
•
2
•
1
Datasets:
text1
stringlengths 7
2.5k
| labels
stringlengths 9
100
|
|---|---|
The command processing function starts by substituting the main module name and path in the hosting process PEB, with the one of the default internet browser. The path of the main browser of the workstation is obtained by reading the registry value | ['T1057'] |
Along the way, HermeticWiper’s more mundane operations provide us with further IOCs to monitor for. These include the momentary creation of the abused driver as well as a system service. It also modifies several registry keys, including setting the SYSTEM\CurrentControlSet\Control\CrashControl CrashDumpEnabled key to 0, effectively disabling crash dumps before the abused driver’s execution starts | ['T1569.002'] |
These Microsoft Office templates are hosted on a command and control server and the downloaded link is embedded in the first stage malicious document | ['T1584.004'] |
Additionally, the IP 211[.]72 [.]242[.]120 is one of the hosts for the domain microsoftmse[.]com, which has been used by several KIVARS variants | ['T1056.001', 'T1113'] |
When communicating with its C2 server, Psylo will use HTTPS with a unique user-agent of (notice the lack of a space between "5.0" and "(Windows | ['T1071.001'] |
In older versions, Valak downloads the second stage JS and uses only one obfuscation technique: Base64. The newer versions use XOR in addition to Base64 | ['T1027'] |
We attribute this activity to TEMP.Zagros (reported by Palo Alto Networks and Trend Micro as MuddyWater), an Iran-nexus actor that has been active since at least May 2017. This actor has engaged in prolific spear phishing of government and defense entities in Central and Southwest Asia. When successfully executed, the malicious documents install a backdoor we track as POWERSTATS | ['T1218.005', 'T1059.005'] |
dlpumgr32.exe, a legitimate signed file that belongs to the DESlock+ product - DLPPREM32.DLL, a malicious DLL sideloaded by dlpumgr32.exe that loads and decodes DLPPREM32.bin - DLPPREM32.bin, a shellcode that decompresses and loads a launcher in memory - data.res, an encrypted file decoded by the launcher and contains two SysUpdate versions: one for a 32-bit architecture and another for a 64-bit architecture - config.res, an encrypted file decoded by the launcher and contains the SysUpdate configuration, such as the command-and-control (C&C) address | ['T1027', 'T1082'] |
The malware has specific features that allow the attackers to perform operations related to online banking transactions, password stealing and clipboard monitoring. We also found various versions of the payload: the version focused on stealing data from victims in Brazil is typically unpacked, while the versions targeting banks in Chile and Mexico are packed with VMProtect or Themida | ['T1027.002'] |
The size of the image is more than 600KB and embedded in it is the encrypted IcedID main module. The encryption algorithm is RC4 and the keys are also embedded in the image at specific offset | ['T1027.003'] |
It is worth noting that in 2019, this actor used a fake file extension (*.png) for the MSI binary hosted on the attacker-controlled GitHub account | ['T1583.006'] |
These variants include system information collection (operating system, computer name), keylogger output, and browser password collection from Internet Explorer, Chrome and Firefox | ['T1082'] |
While Kimsuky is very active, the KONNI RAT has also been upgraded to a more evasive piece of malware | ['T1027.002'] |
But first: How did they get the tools on the victim’s systems. The adversary copied those tools over SMB from compromised system to compromised system wherever they needed these tools | ['T1570'] |
This will also force the victim to re-open the browser using the newly written .lnk file, which is now loaded with Grandoreiro’s malicious extension. This extension will load on every browser startup using this specific .lnk file | ['T1547.001', 'T1036.005'] |
Endpoint Protection . The Trojan.Hydraq Incident . It has been about a week since news of the mysterious Hydraq Trojan (also known as Aurora) attack broke with the unveiling of a threat by Google to pull its operations out of China. Although concrete details of the attacks are not yet public, Google made reference to a number of Gmail accounts that were compromised during or after the attacks. Anatomy of the Attack For a number of years targeted attacks have nearly always followed the same modus operandi. In the more sophisticated attacks, the attacker will use a new zero day vulnerability, as obviously this will have a greater success rate. In this attack a PDF file was used to exploit the Adobe Acrobat, Reader, and Flash Player Remote Code Execution Vulnerability (CVE-2009-1862/BID35759). This PDF installed a Trojan horse which was an earlier version of the current Trojan.Hydraq. This means the remote attacker has the ability to see in real time any user interface activity as if they were sitting right next to the user. As described in the previously posted blog (Hydraq - An Attack of Mythical Proportions), an unpatched Internet Explorer vulnerability (BID 37815) was used as one of the propagation vectors for this particular Trojan.Hydraq attack. This security hole allows remote exploitation, which means that attackers can run any malicious code of their liking on a victim’s machine by taking advantage of the vulnerability. The number of computers we have observed being attacked or have been attacked is low as borne out by our field detection statistics. Prevention & Mitigation Trojan.Hydraq has been known to be spread through specially crafted PDF files and also through malicious Web sites | ['T1005'] |
dbghelp.dll is incompatible with DEP (Data Exception Prevention), as shown in Figure 14. Thus, when it loads the operating system will disable DEP for the injected wmplayer.exe process. This means that code can be executed from memory regions that are not marked as executable in the context of this process | ['T1562.001'] |
For each enumeration, it performs a breadth-first search to wipe the files in the logical drive while ignoring files located in the "%HOMEDRIVE%\Windows" directory. It also only wipes files that have specific file extensions | ['T1083'] |
It is distributed as a set of scripts and encrypted files and utilizes a PowerShell loader based on the Invoke-ReflectivePEInjection PowerSploit module to decode and inject the final payload DLL into memory | ['T1055.001', 'T1055.001', 'T1027'] |
Registry traversal for Putty data exfiltration (left), code showing hostname, username and Private Key Files (right | ['T1552.002'] |
While PotPlayerDB.dat is a variant of PlugX malware, TA416 has updated the payload by changing both its encoding method and expanding the payload’s configuration capabilities. Historically, TA416 relied on the DLL launcher to decode the PlugX payload utilizing an XOR key included at the offset 0 within the PlugX DAT configuration file. One of the main ways it does this is by resolving API functions during runtime. This iteration of PlugX does standard API hashing, but only to resolve the address of the functions GetProcAddress as well as LoadLibrary. Once those functions are resolved properly, it loads the rest of the functions via their text name | ['T1106'] |
After Tor is up and running, Siloscape uses it to connect to its C2 – an IRC server, using an onion address that was provided as a command line argument | ['T1071'] |
One interesting thing to note is that the Keybase account used by the attacker to chat with their victims has the same logo of the Pay2Key EOSIO smart contract system | ['T1585'] |
It should be noted that the Win32/KillDisk.NBB variant used against media companies is more focused on destroying various types of files and documents. It has a long list of file extensions that it tries to overwrite and delete. The complete list contains more than 4000 file extensions | ['T1485'] |
The LNK file contains a list of commands that will be executed upon running, and a blob that is a base64 encoded compressed payload | ['T1027'] |
Fourth, this Darkhotel event is not based on the network protocol C2, but based on a custom file transfer control instruction | ['T1135'] |
When PowerShell is invoked whether via WMI, wscript.exe, or mshta.exe, it executes a one-liner PowerShell code (as outlined above) that reads the encoded text file dropped in ProgramData and then decodes it. The resulting code has multiple layers of obfuscation | ['T1047', 'T1218.005', 'T1559.001', 'T1027'] |
The threat used valid accounts against remote services: Cloud-based applications utilizing federated authentication protocols. Our incident responders analysed the credentials used by the adversary and the traces of the intrusion in log files. They uncovered an obvious overlap in the credentials used by this threat and the presence of those same accounts in previously breached databases. Besides that, the traces in log files showed more than usual login attempts with a username formatted as email address, e.g. username>@<email domain>. While usernames for legitimate logins at the victim’s network were generally formatted like <domain>\<username>. And attempted logins came from a relative small set of IP-addresses | ['T1016', 'T1133'] |
CookieMiner reports all the wallet-related file paths to its remote server so it can later upload the files according to the C2 commands. These files usually include private keys of cryptocurrency wallets. If the victims use iTunes to backup files from iPhone to Mac (can be via Wi-Fi), their iPhone text messages (SMSFILE) will also be retrieved by the attackers (Figure 5 | ['T1083'] |
The JavaScript component is the first stage of the attack and can deploy other malware such as a C# spy component, Golden Chickens components or several Python-based tools. The name Evilnum was given to the C# component by other researchers in the past, but the JS component also has been referred to as Evilnum. We have named the group Evilnum as that is the name of their flagship malware, and we’ll refer to the various malware pieces as components | ['T1105'] |
But Ryuk isn’t new to us… we’ve been tracking it for years. More important than just looking at Ryuk ransomware itself, though, is looking at the operators behind it and their tactics, techniques, and procedures (TTPs)—especially those used before they encrypt any data. The operators of Ryuk ransomware are known by different names in the community, including “WIZARD SPIDER,” “UNC1878,” and “Team9. The malware they use has included TrickBot, Anchor, Bazar, Ryuk, and others | ['T1047', 'T1018'] |
The plugin is executed by using the Info command in the Lizar client application. A data structure containing the OS version, user name and computer name is sent to the server | ['T1082'] |
Observed GoldMax C2 domains are high-reputation and high-prevalence, often acquired from domain resellers so that Whois records retain the creation date from their previous registration, or domains that may have been compromised. This tactic complements NOBELIUM’s operational security strategy as these domains are more likely to be overlooked by security products and analysts based on their perceived long-lived domain ownership. Put simply, several domains we have shared as GoldMax C2 domains are only associated with NOBELIUM after the time they were re-sold or compromised – and Microsoft has provided that indicator context where it is available to us | ['T1584.001'] |
The malware proceeds to blacklist certain processes such as “wininit.exe” when approaches memory scraping in order to speed necessary card scan logic | ['T1057'] |
We were able to collect over fifty samples of the tools used by the Magic Hound campaign using the AutoFocus threat intelligence tool. The earliest malware sample we were able to collect had a compile timestamp in May 2016. The samples themselves ranged from IRC bots, an open source Python remote access tool, malicious macros, and others | ['T1083'] |
From the attacks observed by Volexity, what is most notable is that Patchwork has pivoted its targeting and has launched attacks directly against US-based think tanks. Volexity has also found that, in addition to sending malware lures, the Patchwork threat actors are leveraging unique tracking links in their e-mails for the purpose of identifying which recipients opened their e-mail messages. Strangely, in one case, the threat actors also appear to have used a domain name similar to the Foreign Policy Research Institute (FPRI) in a message purporting to be from CFR. If the exploit is successful, the threat actors will attempt to drop and execute QuasarRAT. Its called the "packager trick" because any file embedded in an RTF file using packager will be automatically dropped to the %tmp% folder (c:\Users\%username%\AppData\Local\Temp) when the RTF document is opened. Second, the threat actors exploit CVE-2017-8570 to achieve code execution via a malicious "scriptlet" file, or .sct file, which is also embedded in the malicious RTF document. The contents of the malicious scriptlet file (displayed below) clearly show the threat actor executing the initial "qrat.exe" dropper from the current user's %tmp% directory. The file, named Microsoft.Win32.TaskScheduler.dll, is digitally signed by a certificate from AirVPN. Conclusion . The addition of US-based think tanks to the list of organizations in the crosshairs of Patchwork shows an increasing diversity in the geographic regions being targeted. Volexity is actively tracking this group and the infrastructure currently in use for the benefit of its network security monitoring and threat intelligence customers | ['T1189'] |
Apply the Microsoft security updates for MS17-010, including the updates for the Windows XP and Windows Server 2003 legacy operating systems. Disable SMBv1 on systems where it is not necessary (e.g. hosts that do not need to communicate with Windows XP and Windows 2000 systems). Carefully evaluate the need for allowing SMBv1-capable systems on interconnected networks compared to the associated risks. Scan networks for the presence of the DoublePulsar backdoor using plugins for tools such as Nmap. Use network auditing tools to scan networks for hosts that are vulnerable to the vulnerabilities described in MS17-010. Implement a backup strategy that includes storing data using offline backup media. Backups to locally connected, network-attached, or cloud-based storage are often insufficient because ransomware frequently accesses and encrypts files stored on these systems | ['T1490'] |
SUPERNOVA is implemented as a modification to the existing ‘app_web_logoimagehandler.ashx.b6031896.dll’ module of the SolarWinds Orion application. The purpose of this module, in it’s legitimate form, is to return the logo image configured by the user to various web pages of the SolarWinds Orion web application. In legitimate operation, this class only contains the ProcessRequest() and LogoImageHandler() methods, a private static Log object, and public boolean parameter IsReusable | ['T1036.005'] |
AgentTesla is a .Net-based infostealer that has the capability to steal data from different applications on victim machines, such as browsers, FTP clients, and file downloaders. One of the new modules that has been added to this malware is the capability to steal WiFi profiles | ['T1555'] |
TA551 has distributed different families of malware, including Ursnif (Gozi/ISFB), Valak and IcedID. TA551 malspam spoofs legitimate email chains based on data retrieved from previously infected Windows hosts. This is a generic statement asking the recipient to open an attached ZIP archive using the supplied password. For example, if the spoofed sender is [email protected], the ZIP attachment would be named companyname.zip. In 2020, we also started seeing emails with info.zip or request.zip as the attached ZIP archive names. These password-protected ZIP attachments contain a Word document with macros to install malware. File names for the extracted Word documents follow noticeable patterns that have evolved as this campaign has progressed. URLs generated by the associated Word macros also follow noticeable patterns that have also evolved as this campaign has progressed | ['T1204.002'] |
The plugins are variously designed to load other tools like Mimikatz or Carbanak, retrieve information from the victim machine, take screenshots, harvest credentials, retrieve browser histories, and more | ['T1217'] |
During our analysis, we successfully extracted the command line argument to execute its payload. The following command is used to execute the payload | ['T1574.002'] |
When receiving HTTP commands, the WellMess server is setup to receive POST requests that contain RC6 encrypted cookies. The server decrypts the cookies using a hardcoded RC6 key and expects the decrypted data to contain no more than four tags | ['T1140', 'T1573.001'] |
Despite the notion that modern cybersecurity protocols have stopped email-based attacks, email continues to be one of the primary attack vectors for malicious actors — both for widespread and targeted operations. Recently, Cisco Talos has observed numerous email-based attacks that are spreading malware to users at both a large and small scale. In this blog post, we analyze several of those campaigns and their tactics, techniques and procedures (TTPs). These campaigns were all observed between mid-May and early July of this year, and can likely be attributed to one, or possibly two, groups. The attacks have become more sophisticated, and have evolved to evade detection on a continual basis. Other researchers have attributed these attacks to a group known as the Cobalt Gang, which has continued its activities even after the arrest of its alleged leader in Spain this year. Simple campaigns typically use a single technique and often embed the final executable payload into the exploit document. The emails either contain a URL pointing to one of the three document types or have initial attack stages attached outright | ['T1059.007'] |
After an initial dormant period of up to two weeks, it retrieves and executes commands, called “Jobs”, that include the ability to transfer files, execute files, profile the system, reboot the machine, and disable system services. The backdoor uses multiple obfuscated blocklists to identify forensic and anti-virus tools running as processes, services, and drivers. TEARDROP and BEACON Malware Used . Multiple SUNBURST samples have been recovered, delivering different payloads. Next it checks that HKU\SOFTWARE\Microsoft\CTF exists, decodes an embedded payload using a custom rolling XOR algorithm and manually loads into memory an embedded payload using a custom PE-like file format. The credentials used for lateral movement were always different from those used for remote access. Detection Opportunity . Organizations can use HX’s LogonTracker module to graph all logon activity and analyze systems displaying a one-to-many relationship between source systems and accounts. After an initial dormant period of up to two weeks, it retrieves and executes commands, called “Jobs”, that include the ability to transfer and execute files, profile the system, and disable system services. The userID is encoded via a custom XOR scheme after the MD5 is calculated. Commands are extracted from HTTP response bodies by searching for HEX strings using the following regular expression: "\{[0-9a-f-]{36}\}"||"[0-9a-f]{32}"||"[0-9a-f]{16}". Command data is spread across multiple strings that are disguised as GUID and HEX strings. The extracted message is single-byte XOR decoded using the first byte of the message, and this is then DEFLATE decompressed | ['T1027'] |
IcedID uses TLS in all of its communication but the certificate is self-signed. They can be spotted, as they use this kind of a self-signed certificate. The keyword “Internet Widgits Pty Ltd” is also being used by Trickbot, another banking malware, and it is believed that Trickbot and IcedID are cousins | ['T1573.002', 'T1185'] |
In December 2019, another version of the AppleJeus malware was identified on Twitter by a cybersecurity company based on many similarities to the original AppleJeus malware. In contrast, open-source reporting stated that the Windows version might have been downloaded via instant messaging service Telegram, as it was found in a “Telegram Downloads” folder on an unnamed victim | ['T1566.002'] |
PowerSploit can be used as a tool for the discovery of stored credentials. Specifically it supports the following modules which will check for credentials encrypted or plain-text in various files and in the registry | ['T1552.002'] |
MegaCortex v1 was executed manually by threat actors using a separate batch file to kill security processes and stop/disable services related to security, backup and shadow copies. That same batch file was subsequently used to execute the MegaCortex binary with a Base64 key as a command-line argument | ['T1489', 'T1562.001'] |
Capable of stealing documents sent to the printer queue. Data gathered for victim recon includes the backup list for Apple mobile devices. Steals written CD images. Capable of stealing files previously seen on removable drives once they are available again. Steals Internet Explorer, Netscape Navigator, FireFox and RealNetworks cookies | ['T1005'] |
First observed by Microsoft on Jan. 13, 2022, WhisperGate malware is computer network attack (CNA) malware aimed at deleting Microsoft Windows Defender and corrupting files on the target. It consists of two samples: One appears as ransomware while the other is a beaconing implant used to deliver an in-memory Microsoft Intermediate Language (MSIL) payload. At the time of writing, there are two known samples identified as WhisperGate: Stage1.exe and Stage2.exe. Stage1.exe purports to be ransomware, as it overwrites the target’s master boot record with 512 bytes and upon reboot displays the following ransom note | ['T1561.002'] |
The second family of Lazarus malware appearing in recent months has, as far as we are aware, received little to no analysis from researchers, possibly due to its targeted nature and a lack of ITW sightings | ['T1105'] |
These cookies are necessary for the website to function and cannot be switched off in our systems. They are usually only set in response to actions made by you which amount to a request for services, such as setting your privacy preferences, logging in or filling in forms. You can set your browser to block or alert you about these cookies, but some parts of the site will not then work. These cookies do not store any personally identifiable information | ['T1082'] |
Of the tools listed above, many were obfuscated with VMProtect (v1.60-2.05), a recurring theme with BackdoorDiplomacy tools | ['T1027'] |
An uptick in activity from GRIM SPIDER, a subgroup of the criminal enterprise CrowdStrike Intelligence tracks as WIZARD SPIDER, has led to the identification of consistent actions employed to carry out their attacks. As part of their initial compromise — usually as a download from a spam email — they gain a foothold with their modular TrickBot malware, which was developed and is principally operated by WIZARD SPIDER. Once TrickBot is executed, new enumeration modules are downloaded onto the compromised machine to facilitate WIZARD SPIDER’s spread in search of credentials with the aim of gaining access to the domain controller. The criminal actors use RDP to perform lateral movement and explore the victim environment, with an end result of gaining access to the domain controller. Once this access has been achieved, GRIM SPIDER is able to deploy the Ryuk ransomware to the entire network | ['T1071.001', 'T1021.001', 'T1204.002', 'T1041'] |
Unlike recent variants of Mirai and Gafgyt that target vulnerable Linux systems via randomly generated IP addresses, Xbash also scans and trawls through domain names. The C&C scans for specific destinations’ known vulnerabilities in Hadoop, Redis and ActiveMQ (CVE-2016-3088) for self-propagation. Hadoop’s unauthenticated command execution flaw discovered in October 2016, as well as the Redis arbitrary and remote command execution vulnerability disclosed in October 2015, have yet to be assigned CVE numbers. Based on the active C&C traffic, it scans and probes for open TCP or UDP ports such as HTTP, VNC, MySQL/MariaDB, Telnet, FTP, MongoDB, RDP, ElasticSearch, Oracle Database, CouchDB, Rlogin and PostgreSQL. While the malware uses a weak username and password dictionary to brute force itself into the service, it is also able to update its set from the C&C server, delete all the databases, and display the ransom message | ['T1203'] |
As can be seen in the figure above, the packer used for CVE-2019-0803 is very similar to the one used in CVE-2017-0005. The file was compiled on September 18, 2018, and is also internally named “Add.dll”. Like the previously packed exploit, CVE-2019-0803 also has an export function named “AddByGod” and contains debug information | ['T1027.002'] |
Although, the use of target names with actuating themes is not new to this group, there has been a significant uptick in the number of emails received and this campaign has been persistently active for the past few weeks | ['T1566.002'] |
This function is the supporting functionality for WinVNC. To allow the VNC session to connect, the current network socket WSAProtcol_Info structure is written to a named pipe prior to calling zxFunction001 | ['T1021.005'] |
We also analyzed further Gamaredon tools that have the ability to inject malicious macros and remote templates into existing Office documents. Tools linked to Gamaredon and discussed in this blogpost are detected as variants of MSIL/Pterodo, Win32/Pterodo or Win64/Pterodo by ESET’s products. Contrary to other APT groups, the Gamaredon group seems to make no effort in trying to stay under the radar. It also saves to disk the malicious OTM file (Outlook VBA project) that contains a macro, the malicious email attachment and, in some cases, a list of recipients that the emails should be sent to. Office macro injection module – CodeBuilder . We analyzed different variants of malicious modules used by the Gamaredon group to inject malicious macros or remote templates into documents already present on the compromised system. Module updates . Interestingly, some of the custom tools described in Palo Alto Networks’ 2017 blogpost on Gamaredon are still being updated and in use today. C# compiler module . This .NET executable, similar to many other tools used by the Gamaredon group, uses obfuscation techniques such as junk code insertion and string obfuscation. As with many other tools used by the Gamaredon group, they come in four different coding languages: C/C++, C#, batch file and VBScript. Quality of execution . We were able to collect numerous different samples of malicious scripts, executables and documents used by the Gamaredon group throughout their campaigns. Conclusion . Despite the simplicity of most of their tools, the Gamaredon group also is capable of deploying some novelty, such as their Outlook VBA module | ['T1547.001'] |
b.wnry — Bitmap image used as desktop wallpaper (shown in Figure 2) - c.wnry — Configuration containing Tor command and control (C2) addresses, Bitcoin addresses, and other data - r.wnry — Ransom demand text - s.wnry — ZIP archive containing Tor software to be installed on the victim’s system; saved in TaskData directory - t.wnry — Encrypted DLL containing file-encryption functionality - u.wnry — Main module of the WCry ransomware “decryptor” - taskdl.exe — WNCRYT temporary file cleanup program - taskse.exe — Program that displays decryptor window to RDP sessions - msg — Directory containing Rich Text Format (RTF) ransom demands in multiple languages | ['T1090.003'] |
CISA also observed the threat actor using open-source tools such as Plink and TightVNC for lateral movement. CISA observed the threat actor using the techniques identified in table 8 for lateral movement within the victim environment | ['T1090', 'T1021.004', 'T1572'] |
In this campaign, the group sent spear phishing emails containing malicious documents that led to the installation of the UPPERCUT backdoor. Part of this blog post will discuss the updates and differences we have observed across multiple versions of this backdoor | ['T1204.002'] |
After the ransomware is executed, Clop appends the .clop extension to the victim's files. We have observed different variants using different extensions, such as “.CIIp”, “.Cllp” and “.C_L_O_P”. Different versions of the ransom note have also been observed after encryption. Depending on the variant, any of these ransom text files could drop: “ClopReadMe.txt”, “README_README.txt”, “Cl0pReadMe.txt“ and “READ_ME_!!!.TXT | ['T1486'] |
H1N1 has self-propagation/lateral movement functionality (which requires user interaction) via mapped/available network shares or mounted USB devices | ['T1080'] |
Another component of the KGH suite is the m.dll module, which is an information stealer that harvest data from browsers, Windows Credential Manager, WINSCP and mail clients | ['T1114.001'] |
Before the driver is loaded, the malware disables crash dump by setting the following registry key | ['T1070', 'T1562.006', 'T1112'] |
Conclusion Tick has left a trail of evidence indicating that its activity began as early as 2006. In earlier attacks, the group used malicious Microsoft Word documents to infect victims, with compromised websites being added to the mix as a more recent attack vector | ['T1204.002'] |
Gathering system information and sending it to the control server. The system information gathered from the endpoint includes: MAC address of the endpoint Computer Name Product name from HKLM\Software\Microsoft\Windows NT\CurrentVersion ProductName This information is concatenated into a single string in the format: “MAC_Address||ComputerName||ProductName” and is sent to the control server - MAC address of the endpoint - Computer Name - Product name from HKLM\Software\Microsoft\Windows NT\CurrentVersion ProductName - This information is concatenated into a single string in the format: “MAC_Address||ComputerName||ProductName” and is sent to the control server - Recording HTTP requests from the control server to the temporary file prx in the implant’s install directory with the current system timestamp | ['T1012'] |
Credential hopping for obscuring lateral movement - Office 365 (O365) Service Principal and Application hijacking, impersonation and manipulation - Stealing browser cookies for bypassing multifactor authentication - Use of the TrailBlazer implant and the Linux variant of GoldMax malware - Credential theft using Get-ADReplAccount | ['T1078.002', 'T1550.001'] |
In the analyzed sample the RAT component was named “BotDLL[.]dll”. It has some typical RAT functionality such as command shell, video recording of the screen, remote desktop, port forwarding, and file system access | ['T1125', 'T1090', 'T1005'] |
This final cluster appears to serve as the C2 infrastructure for a custom remote administration tool called Pteranodon. It is capable of downloading and executing files, capturing screenshots and executing arbitrary commands on compromised systems | ['T1113'] |
Introduced in macOS 10.11, this utility has only one publicly documented use, which is to return the status of the System Integrity Protection tool. The csrutil tool is commonly used by malware and post-exploitation tools to determine whether certain files and directories on the system are writable or not | ['T1082'] |
Shortly after this RTF document is opened, the remaining stages of the Inception malware are found executing on the system. The loader DLL is responsible for decrypting and injecting the core payload DLL into memory, from an encrypted file present on disk. The core payload DLL's main function is to gather system information, execute other malware in the form of plugins, and update itself | ['T1204.002'] |
The NOKKI malware itself has been updated in the short period of time it has been observed, moving from FTP to HTTP for C2 operations. The malware is modular in nature, and based on analysis of the information gathering module, it is highly likely the NOKKI operators are the same as the KONNI operators | ['T1071.001', 'T1071.002'] |
The network mode being set to the host along with the container trying to be deployed as a privileged container. The Docker Hub account of MegawebMaster has numerous public images, five of which have TeamTNT utilities with a significant amount of downloads. These five images include dockgeddon, docker, tornadopw, and dcounter (T1204.003 | ['T1496'] |
TeamTNT has also been spotted using a malicious Docker image which can be found on Docker Hub to infect its victims’ servers | ['T1610', 'T1071.001'] |
Even simple API calls were obfuscated, and instead of just calling the functions, Siloscape made the effort to use the Native API (NTAPI) version of the same function | ['T1106', 'T1027'] |
The DLL expects the export named 'Add' to be used when initially loaded. When this function is executed PLAINTEE executes the following command in a new process to add persistence | ['T1059.003', 'T1547.001'] |
In this version, the communication protocol with the C&C server was also upgraded to use AES encryption | ['T1573.001'] |
The domain name is generated based on the current month and year values, e.g. for August 2017 the domain name used would be “nylalobghyhirgh.com | ['T1568.002'] |
HAYMAKER is a backdoor that can download and execute additional payloads in the form of modules. It communicates encoded system information to a single hard coded command and control (C2) server, using the system’s default User-Agent string. BUGJUICE is a backdoor that is executed by launching a benign file and then hijacking the search order to load a malicious dll into it. It has the capability to find files, enumerate drives, exfiltrate data, take screenshots and provide a reverse shell. SNUGRIDE is a backdoor that communicates with its C2 server through HTTP requests. The malware’s capabilities include taking a system survey, access to the filesystem, executing commands and a reverse shell | ['T1059.003'] |
The attackers then attempt to gain root access to the server by setting up a local privileged user named ‘hilde’ on the host server and use it in order to connect back via SSH | ['T1021.004'] |
The Russia-linked Shuckworm group (aka Gamaredon, Armageddon) is continuing to conduct cyber-espionage attacks against targets in Ukraine. Over the course of recent months, Symantec’s Threat Hunter Team, a part of Broadcom Software, has found evidence of attempted attacks against a number of organizations in the country | ['T1057', 'T1204.002'] |
After the malware has invoked a method named _s_is_high_time and waited on several timers to expire, it begins encrypting the (unfortunate) user’s files, by invoking a function named carve_target. It then generates a list of files to encrypt, by invoking the get_targets function, passing in the is_file_target as a filter function. This filter function filters out all files, except those that match certain file extensions. The encrypted list of extensions is hard-coded at address 000000010001299E within the malware. In part one of this blog post series, we decrypted all the embedded string, thus can readily examine the decrypted list | ['T1486'] |
Timeline . OSX/FruitFly: 1) 2) Remove the malicious launch agent plist file ~/Library/LaunchAgents/com.client.client.plist 3) Remove the malware's persistent perl script & file. Ok, so the attackers are using an open-source multi-stage post-exploitation agent. Unfortunately this file is now inaccessible. The author of the thread announced a RAT dubbed Proton, intended for installation exclusively on MAC OS devices. Finally, the malware modifies the infected host's network settings in order to set up a proxy who's address is (dynamically) specified via a remote proxy auto-configuration (PAC) file. As it's a binary plist file, dump its contents with the plutil utility (using the -p commandline flag): . As the KeepAlive key has been set to 1 (true), the Launch Daemon will be automatically started everytime the infected system is rebooted. MacRansom is the the first 'Ransomware-as-a-Service' for macOS, that aims to encrypt (ransom) all user's files. Then these files will be passed (to a new instance) of the malware, in order to be encrypted. Thus it appears that once encrypted, the files are pretty much gone for good (save for a perhaps a brute force decryption attack). Good news, RansomWhere. Using the neat 'Suspicious Package' application, we can statically examine this script: In short, it persists CPUMeaner as a launch agent via the /Library/LaunchAgents/com.osxext.cpucooler.plist file | ['T1140'] |
Various scans and queries are used to find proxy settings, domain controllers, remote desktop services, Citrix services, and network shares. Otherwise, a jump host or other system likely used by domain admins is found and equipped with a Cobalt Strike beacon | ['T1012'] |
This attack begins with a spear phishing attack through a targeted email campaign. Over 80 files were sent to 40 email accounts within the organization, within the span of about an hour. The email contains Microsoft Excel attachments with malicious macros | ['T1566.001'] |
The original Microsoft Excel spreadsheet is copied into the %TEMP% directory - The embedded object “xl\embeddings\oleObject1[.]bin” inside the Microsoft Excel spreadsheet is copied into the %TEMP% directory - The DLL inside oleObject1.bin is extracted and copied into %APPDATA% by the “ReadAndWriteExtractedBinFile” function - The DLL is loaded with LoadLibraryA - The DLL’s exported function, such as “Get2”, is run by the macro | ['T1055.001'] |
Please note that the Ecipekac Layer III loader module is embedded in the encrypted Layer II loader | ['T1027'] |
This feature generates a stageless Beacon payload artifact, hosts it on Cobalt Strike’s web server, and presents a one-liner to download and run the artifact | ['T1197'] |
The NetWire payloads in all observed campaigns included nearly identical configurations. Specifically, the C2 domain clients[.]enigmasolutions[.]xyz and the password were the same | ['T1105'] |
Upon exploitation, a GH0ST RAT variant is delivered to the victims’ system, which calls out to a previously known APT18 CnC address 223.25.233.248. GH0ST RAT is a backdoor derived from public source code | ['T1070.001', 'T1059'] |
If the attack progresses, the user will be taken to the download of an MS Word document containing malicious macros that has very low detection rate at the moment of this campaign delivery. From a metadata standpoint, the document does not include any specific signal or characteristic that would help us tracking documents from the same author, as shown in Figure 6 | ['T1566.001', 'T1059.005'] |
Each Casbaneiro sample using this method has the buyer’s ID hardcoded in its data. When it downloads such configuration file, it parses it and finds the line that is intended for the specific buyer’s ID and downloads and executes the payload | ['T1547.001', 'T1036.005', 'T1105'] |
However, while the malware used in these new attacks uses similar infection mechanisms to PlugX, it is a completely new tool with its own specific behavior patterns and architecture. We have named this tool “BBSRAT. Targeting and Infrastructure . As described in earlier reports on “Roaming Tiger”, the attack observed in August 2015 used weaponized exploit documents that leave Russian language decoy document files after infecting the system. Figure 2 confirms that the decoy document that opens after the malware infects the system is indeed a list of international exhibitions that were conducted on Russian territory in 2015. Analysis of the command and control (C2) infrastructure shows that the newly discovered samples of BBSRAT used the same C2 domains as previously published in the “Roaming Tiger” campaign, including transactiona[.]com and futuresgold[.]com. This may indicate that for the newer attack campaign using BBSRAT, the adversary may have deployed purpose-built variants and/or infrastructure for each of the intended targets. As we can see, the second command is specifically crafted to run on 64-bit versions of Microsoft Windows. Every subsequent request made by BBSRAT increments this counter by one. The following commands and sub-commands have been identified: Please refer to the appendix for a full list of identified BBSRAT samples and their associated C2 servers. Despite the fact that the information about these attackers has been public for over a year, including a listing of many of the command and control servers, they continue to reuse much of their exposed playbook | ['T1546.015'] |
This DLL has no other noticeable characteristics, as it functions like a typical malicious sideload. After loading the encrypted payload in memory, it transfers the execution to a shellcode that is located at the beginning of the file. Once loaded in memory, the ZeroT shellcode does not present any kind of obfuscation, unlike that for PlugX. This shellcode is charged with unpacking the encrypted and compressed payload. As in the new PlugX dropper detailed below, this is done using RC4 and RtlDecompressBuffer. As in PlugX samples, the PE header of ZeroT has been tampered with, specifically the “MZ” and “PE” constants (Fig | ['T1573.001'] |
All trusted domains, domains, and domain controllers - A list of computers and network devices on the network - The infected machine user and groups the user belongs to - The infected machine, including machine name, operating system, workstation domain, and more information - Network adapters that have connected to the machine and DNS servers | ['T1069', 'T1033', 'T1016'] |
So this method uses psexec itself to copy the payload over the network, overwrite earlier versions (if found), and run it without waiting for any response | ['T1570'] |
Sleeps the downloader. After that, it downloads a file from Discord. The downloaded file is in reverse byte order. Downloads file from Discord. The downloader restores the downloaded file by reversing the bytes within the file. Method that reverses the downloaded file. The restored file is a DLL and serves as the third stage of the infection chain. Retrieving third-stage public methods using Type.GetMethods | ['T1105', 'T1027'] |
End of preview. Expand
in Dataset Viewer.
The Security Attack Pattern (TTP) Recognition or Mapping Task
We share in this repo the MITRE ATT&CK mapping datasets, with training, validation and test splits.
The datasets can be considered as an emerging and challenging multilabel classification NLP task, with over 600 hierarchical classes.
NOTE: due to their security nature, these datasets contain textual information about malware and other security aspects.
Datasets
TRAM
This dataset belongs to CTID, is originally provided in this github link.
We processed the original files (i.e., gather from all sources, remove duplicates, resolve noisy / too short text and noisy labels, remap to MITRE ATTACK 12.0) and split into training, dev and test splits.
Procedure+
The dataset consists of two sub- datasets:
- Procedures: belong to MITRE. All procedure examples from v12.0 are gathered and processed (i.e., remove markups) and split into training, dev and test splits.
- Derived procedures: we crawled the URL references for each procedure example, and extract original text from the articles that are determined to be relevant to the procedure examples. The text are processed and split into training, dev and test splits.
Expert
The dataset is constructed from a large pool of high-quality threat reports.
The rich textual paragraphs are carefully selected and then annotated by seasoned security experts.
The dataset is also pre-split into training, dev and test splits. There are ~4 labels per text in the test split, on average.
Citations
If you use the datasets in your research or want to refer to our work, please cite:
@inproceedings{nguyen-srndic-neth-ttpm,
title = "Noise Contrastive Estimation-based Matching Framework for Low-resource Security Attack Pattern Recognition",
author = "Nguyen, Tu and Šrndić, Nedim and Neth, Alexander",
booktitle = "Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics",
month = mar,
year = "2024",
publisher = "Association for Computational Linguistics",
abstract = "Tactics, Techniques and Procedures (TTPs) represent sophisticated attack patterns in the cybersecurity domain, described encyclopedically in textual knowledge bases. Identifying TTPs in cybersecurity writing, often called TTP mapping, is an important and challenging task. Conventional learning approaches often target the problem in the classical multi-class or multilabel classification setting. This setting hinders the learning ability of the model due to a large number of classes (i.e., TTPs), the inevitable skewness of the label distribution and the complex hierarchical structure of the label space. We formulate the problem in a different learning paradigm, where the assignment of a text to a TTP label is decided by the direct semantic similarity between the two, thus reducing the complexity of competing solely over the large labeling space. To that end, we propose a neural matching architecture with an effective sampling-based learn-to-compare mechanism, facilitating the learning process of the matching model despite constrained resources.",
}
License
This project is licensed under the Creative Commons CC BY License, version 4.0.
- Downloads last month
- 163
Size of downloaded dataset files:
3.77 MB
Size of the auto-converted Parquet files:
2.06 MB
Number of rows:
20,736
Models trained or fine-tuned on tumeteor/Security-TTP-Mapping