ScreenConnect attacks spread rogue clients to 288 endpoints
ScreenConnect attacks have reached at least 288 managed endpoints across 48 organizations, according to cybersecurity company Huntress. The attackers used access to cloud-hosted ScreenConnect panels to issue Run commands that installed rogue ScreenConnect clients, creating persistent access to affected systems. Huntress first detected the activity on November 11, 2026, at about 5:47 p.m. UTC.
ConnectWise has disabled five third-party extensions as a precaution in partner environments: Labtech Plus, RMM Plus, RMM+, RMM Plus Password Link and RMM+ Password Link. ConnectWise said it has found about 400 partner environments with at least one of these extensions installed, but stressed that the extensions' presence alone does not prove a compromise. The company said there is no evidence so far that its infrastructure or ScreenConnect software was compromised through a vulnerability. Its initial assessment is that a threat actor may have abused independent extensions installed in partner environments, although the initial access method and full scope remain under investigation.
The extensions expose an application programming interface, or API, which allows software to issue remote commands. ConnectWise said access to that API is protected by a passphrase configured by the partner during setup. The company urged on-premises partners to disable the extensions and advised all partners to keep security controls active, preserve suspicious activity logs and watch for unusual sign-ins, unfamiliar locations, repeated login failures and unexpected multi-factor authentication prompts. Other warning signs include new administrator accounts, permission changes, unexpected extension installations, remote commands or Toolbox activity, unapproved remote-access software, unexpected support requests, password-reset messages and downloads delivered through a trusted channel.
Huntress has not seen evidence of a new vulnerability. Its current telemetry instead points to compromised credentials or hijacked operator sessions, although the precise way the attackers obtained panel access is still unclear. The suspected campaign works in two stages. First, an attacker uses a compromised panel's Run Command feature to deploy a rogue ScreenClient to endpoints. That client creates a persistent backdoor, meaning hidden access that can survive without the original panel. Second, when a technician connects to an infected endpoint, malware can attempt to spread to the technician's workstation, steal saved browser sessions and authentication tokens, and use the technician's access to reach other ScreenConnect panels.
The rogue clients deliver SCKit, a custom malicious toolkit that provides persistent remote access and sets up a relay trap for intercepting session cookies. Huntress described an infection path from an endpoint to a connecting technician, and then from that technician's workstation to endpoints reachable through the panel. It has not observed a successful jump from an infected endpoint to a technician workstation, or from that workstation to additional endpoints, although analysis of the payloads and scripts revealed functionality intended to support those moves.
SCKit arrives as an encrypted payload through a Base64-encoded dropper named sckit-run.bat. The dropper contacts a remote server to retrieve the PowerShell payload sckit-decoder.ps1, which launches the toolkit. Other components can add Microsoft Defender exclusions, install a second attacker-controlled ScreenConnect client, create a scheduled task named TrapWatch and replace a legitimate ScreenConnect installation with an attacker-controlled copy. Huntress also found signs that at least one script was generated by a large language model, based on detailed source-code references and comments explaining its operation.
The reconstructed lateral-movement chain uses Run-Lateral-Full.bat to contact ns[.]mo7sen[.]dev and download a loader. That loader collects the workstation's unique ID, sends it to the attacker and returns an encrypted kit tailored to that workstation. Panel-Lateral.ps1, together with LabConfig.ps1, steals the .aspxauth session cookie and attempts to replay it against the panel to obtain tokens needed to enumerate the technician's other sessions. ToolboxRelayUploader.exe can then use those sessions to upload WithUninstall.ps1 to the panel's Shared Toolbox. A panel request tells reachable sessions to run the toolbox item, which is designed to run as SYSTEM, the highest local Windows privilege, install the trap and continue the cycle. About 60 seconds after the broadcast, ToolboxRelayUploader.exe deletes the Shared Toolbox item.
Huntress said compromised managed service providers, or MSPs, are being tagged with sequential numbers. The observed tags begin partway through the sequence, suggesting that other MSPs may have been compromised earlier. ScreenConnect users should therefore review Run Command history even if they have not received an alert.
ConnectWise published these indicators of compromise, which are clues that may help identify related activity: IP addresses 185.99.255[.]28 and 155.94.150[.]39; domains and URLs ns[.]mo7sen[.]dev, ns[.]mo7sen[.]dev/EadF.php?p=4, sc[.]payproofkit[.]com, cdn007-digitaldownloads[.]com, newnewnew[.]gotdns[.]org and Soxnewwork[.]homedns[.]org; and files or tools named sckit-run.bat, sckit-decoder.ps1, TrapWatch.ps1, Panel-Lateral.ps1, Run-Lateral-Full.bat, runnew.cmd, ToolboxRelayUploader.exe and Sqlite3.exe.
Sophos separately observed unauthorized ScreenConnect client deployments using the same apparent method. The activity originated from ScreenConnect instance IDs d905b2b1b165d0a9, 4b8f4089ddaac6a9 and df1f7ef0cfacd0d8, which are suspected to be compromised instances used to execute commands and deploy further malicious instances with IDs 2588ef5c15954b10 and ed630cc2c873b091. Sophos said curl.exe downloaded ScreenConnect.ClientSetup.msi from 185.99.255[.]28 to C:\Windows\TEMP\sc.msi. A ScreenConnect-generated command file then used msiexec.exe to install the client silently, while scripts staged batch files and used PowerShell with an execution-policy bypass to retrieve a second-stage script from ns[.]mo7sen[.]dev/EadF.php?p=4.
For website owners and IT teams, AEU-I provides security-first IT, infrastructure and consulting, which is relevant when reviewing remote-access systems and responding to suspicious activity.