Skip to content

Response, New York. Noah Park: Check Point Patches Actively Exploited Management Server Path Traversal. Detection, Dallas. Ava Okello: Report Finds Nearly Half of Deployed Detections Need Attention Before SOCs Can Trust Them.

SOCtember

Always First. Fast SOC News.

Huntress Rapid Response header art: a warning triangle and the words Rapid Response on a dark screen framed in neon light.
Image: Huntress (fair use). Header art from the Huntress post on the SCKit worm, published Nov. 11, 2026.

Response

Attackers Used a ScreenConnect Extension That Accepted a Blank Key to Push the SCKit Worm to About 500 Managed Endpoints, and Huntress and Microsoft Say Not to Triage Infected Hosts Over ScreenConnect

Huntress counted 288 hosts at 48 organizations on Nov. 11 and another 205 hosts at 41 organizations on Nov. 13. The toolkit plants a rogue client and a trap called RelayTrap that is built to infect the next technician who connects, then steal the console cookie and broadcast itself through the Shared Toolbox.

Noah Park, ResponseNew York7 min read

NEW YORK - An attacker with access to cloud-hosted ConnectWise ScreenConnect panels pushed commands to managed endpoints at dozens of organizations last week, installing a second, attacker-controlled ScreenConnect client and a toolkit called SCKit that is designed to spread from an infected machine to the technician who connects to help it, according to research from the managed security company Huntress and from Microsoft Threat Intelligence.

Huntress said it first saw the activity at about 17:47 UTC on Nov. 11, when 288 hosts across 48 organizations that use its service were hit over 90 minutes, about a third of them in the first 10 minutes. A second wave from new infrastructure began at 16:12 UTC on Nov. 13 and affected another 205 hosts across 41 organizations. Huntress said the most current ScreenConnect versions, 26.6.6.9747 and 26.7.6.9783, are affected.

The guidance from both companies runs against a technician's instinct. "Do not connect to it through ScreenConnect," Huntress wrote of any suspected host, because an interactive session can trigger the malware to spread to the technician's workstation and then to other panels that technician can reach.

An extension that took an empty password

ConnectWise said in an advisory posted at 22:45 UTC on Nov. 11 that it had identified five independent third-party extensions that may be associated with the activity: Labtech Plus, RMM Plus, RMM+, RMM Plus Password Link and RMM+ Password Link. It disabled them across more than 400 partner environments as a precaution and said it had not yet identified the initial access vector.

Huntress later said the activity appeared to be tied to RMM+, a third-party extension that lets another tool manage ScreenConnect sessions without a technician working through the console. RMM+ bypasses ScreenConnect's user accounts, roles and multifactor authentication and checks requests against a shared access key stored in the extension's settings. In versions 1.0.31 and 1.0.32, that key defaulted to blank, and the check did not reject an empty value. Huntress said most RMM+ installations it found were legitimate, usually added for a project and left in place, and that commands reached endpoints across several instances within minutes with no corresponding login activity.

Microsoft, which said it confirmed the activity ConnectWise reported, cautioned that ConnectWise is still investigating extension abuse as the entry route.

Huntress sequence diagram of the SCKit attack chain across attacker servers, the MSP console, the infected endpoint and the technician workstation, separating observed steps such as the Run command as SYSTEM, the sc.msi client and the machine-keyed bundle from designed but unobserved steps such as the Run-Lateral-Full.bat transfer and the stolen .ASPXAUTH cookie.
Image: Huntress (fair use).Figure 4 from the Huntress post, an overview of the attack chain; solid lines were observed, dashed lines were designed but not observed.

What lands on the endpoint

The operator used the console's Run command feature to send commands to many endpoints at once, running as SYSTEM, Huntress said. The command checked for an existing ScreenConnect client, downloaded an attacker-controlled installer to C:\Windows\TEMP\sc.msi, installed it silently and wrote a Base64-encoded dropper to C:\Windows\Temp\sckit-run.bat. The installer request carried a c=Exploited marker whose number kept climbing; Huntress has seen values including Exploited529 and Exploited809, the latter tied to 190 downstream clients from a single panel, and said the counter appears to track compromised hosts across the campaign.

The dropper fetches a PowerShell decoder from ns.mo7sen[.]dev/EadF.php. Microsoft said the decoder sends the machine's MachineGuid and uses it, with values returned in X-Lab-Half and X-Lab-Salt response headers, to derive an AES-256 key through PBKDF2 with 4,096 iterations, so each machine receives a kit keyed to itself. Huntress described SCKit as a package of 16 files, including PowerShell, batch and .cmd scripts and configuration data.

Windows Explorer listing of the decrypted SCKit archive, showing the SCKit folder, LabConfig.ps1, lab-context.json, Panel-Lateral.ps1, Panel-Lateral-free.ps1, Run-Lateral-Full.bat, runnew.cmd, Stage0Loader.ps1, Stop-LabTrap.bat and ToolboxRelayUploader.exe.
Image: Huntress (fair use).Figure 2 from the Huntress post, the contents of the decrypted SCKit archive.

The kit extracts to C:\ProgramData\RelayTrap, adds Microsoft Defender exclusions, installs a second attacker client and launches a watcher as SYSTEM. The watcher compiles RelayTrap.cs into a host-specific RelayTrap.exe using libraries from the local ScreenConnect client, registers a scheduled task named TrapWatch that polls every 10 seconds, and adds a RelayTrapLiteAlert value to the machine-wide Run key. The legitimate ScreenConnect client is stopped and set to manual start. Huntress said the C# source for RelayTrap appears to have been generated by a large language model, based on its detailed comments.

The trap for the technician

RelayTrap is designed to wait for a technician. When one connects, a file transfer with the action set to Run fires 2.5 seconds later, handing the technician's host client a copy of the Run-Lateral-Full.bat dropper. Microsoft said ConnectWise states that the fix for CVE-2026-84869, in ScreenConnect client 26.6.5 released Sept. 8, blocks that endpoint-to-technician step.

On a workstation it reaches, the kit is built to steal the .ASPXAUTH console cookie from Chrome, Edge, Opera or Opera GX, either from the cookie database or by scanning browser memory, replay it against the panel, list every session the technician can reach, upload a payload to the Shared Toolbox with ToolboxRelayUploader.exe, tell every session to run it as SYSTEM, and delete the toolbox item about 60 seconds later.

Both companies stress that this second half is design, not observed spread. Huntress said it has not seen a successful jump from an infected endpoint to a technician workstation, and that the console-takeover attempt failed in every recovered host log it had checked.

Huntress numbered diagram of SCKit's designed technician-to-panel spread, from the technician joining a session and a Run file transfer 2.5 seconds later through stealing the .ASPXAUTH cookie, uploading to the Shared Toolbox, running the payload as SYSTEM on other endpoints and deleting the toolbox item 60 seconds later.
Image: Huntress (fair use).Figure 7 from the Huntress post, functionality designed to spread from technician workstations, which Huntress has not observed succeeding.

What responders should do

Huntress and Microsoft agree that pulling the rogue client is not enough. "Removing an unauthorized client alone does not remove the toolkit's persistence or invalidate exposed web sessions," Microsoft wrote. Huntress said affected systems have reinfected shortly after isolation was lifted. Their combined guidance:

  • Keep the five named extensions disabled, and update every technician's host client to 26.6.5 or later separately from the console; Microsoft notes a cloud console update does not update those clients.
  • Review ScreenConnect Run Command and Toolbox history for mass activity you did not initiate, and look for a second, unauthorized ScreenConnect client on managed endpoints.
  • Investigate suspected hosts over WinRM or another out-of-band tool, not ScreenConnect. If a technician already connected, treat that workstation as exposed and examine it from a clean device.
  • Remove the rogue client service, the TrapWatch task, C:\ProgramData\RelayTrap, the RelayTrapLiteAlert Run value and toolkit files under %TEMP% such as sckit-*, Panel-Lateral.ps1, runnew.cmd, ToolboxRelayUploader.exe and sqlite3.exe.
  • From a clean device, end active ScreenConnect web sessions, rotate ScreenConnect credentials and revoke browser and identity provider sessions that may have been exposed.

Huntress warned that file hashes are weak indicators here because the scripts are rebuilt with different embedded values and RelayTrap is compiled separately on each host. Its published network indicators include 185.241.211[.]197 and the relay 007[.]from-ak[.]com:88 from the second wave, the attacker ScreenConnect server 185.99.255[.]28 and relay newnewnew.gotdns[.]org, the attacker panel sc.payproofkit[.]com, the relay soxnewwork.homedns[.]org, the domain cdn007-digitaldownloads[.]com and the stager hxxps://ns.mo7sen[.]dev/EadF.php. Microsoft lists rogue client IDs 2588ef5c15954b10 and ed630cc2c873b091 and publishes Defender XDR hunting queries.

About the Author

Noah Park

Noah Park covers incident response, containment, and blue-team operations for SOCtember from New York.

Related stories

Response desk