Updated November 17, 2026: Huntress has now confirmed in-the-wild exploitation of CVE-2026-84869. The SCKit campaign’s own worm-like spread still failed.

On November 11, attackers broke into cloud-hosted ScreenConnect environments used by managed service providers (MSPs) and used the platform’s own Run command feature to install rogue remote-access clients on customer machines. Within 90 minutes, Huntress counted 288 compromised endpoints across 48 organisations. About a third were hit in the first 10 minutes.

The payload behind the push is a toolkit called SCKit. It is not a single virus. It is a collection of 16 scripts and components built to keep attackers inside a managed environment, and then to turn a routine support session into an attack on the technician on the other end.

That second part is the headline. It is also the part that has not been shown to work. Understanding the difference matters for anyone who runs or buys managed IT services.

What happened on November 11

ScreenConnect, formerly ConnectWise Control, is remote-support software. MSPs use it to connect to clients’ computers, fix problems, and manage devices without visiting an office. The same access that makes support fast makes the platform a valuable target. One compromised console can reach every customer that console manages.

In this campaign, the attackers did not need to break the ScreenConnect server itself. According to Huntress, they got in through RMM+, a third-party extension that lets outside management tools start and control ScreenConnect sessions through an API. Versions 1.0.31 and 1.0.32 had an authentication flaw: if an administrator had never set the extension’s shared access key, the extension accepted a blank one. An attacker could send an empty key and pass the check.

That matches what Huntress saw in the logs: commands ran across multiple ScreenConnect instances without a matching technician login, and the activity was tied to RMM+. ConnectWise has disabled five related extensions across more than 400 partner environments as a precaution. The company says installing an extension does not prove a compromise, and its investigation found no evidence that ConnectWise’s own infrastructure was breached or that the initial access was a ScreenConnect vulnerability. The entry point is still officially under investigation; the extension flaw is the leading explanation, not a closed case.

The deployment worked against current ScreenConnect releases because the attackers were abusing legitimate remote-management features, not exploiting a server bug. Huntress later reported a second spike on November 13, with 205 more hosts across 41 organisations. Blackpoint Cyber said that wave appeared to target on-premises instances where the mitigations were still pending.

This is the same pattern we covered in our earlier report on the ScreenConnect attacks: the management layer is the target, and the extension layer is part of the attack surface.

What is SCKit?

SCKit is the toolkit Huntress and Blackpoint recovered after the November 11 intrusions. Blackpoint’s Adversary Pursuit Group pulled the loader from attacker-controlled infrastructure and decrypted the later stages. The kit includes 16 scripts and components. Its jobs are simple: keep access, hide, and spread.

First, the attackers install additional ScreenConnect Access clients that connect to servers they control. Those clients give them a way back in even if the legitimate MSP session is closed. The kit also adds persistence through scheduled tasks and registry Run entries, creates Defender exclusions, and changes lockout policies. In one recovered chain, a rogue client is installed from a temporary MSI file and a scheduled task named TrapWatch is created to keep the trap running.

None of that is unusual for post-exploitation malware. The unusual part is RelayTrap.

How RelayTrap tries to turn support into a trap

A normal ScreenConnect session has two sides. The technician runs the Host client. The customer’s computer runs the Guest. The technician can send files to the Guest and run them. The Guest is not supposed to be able to push a file back and have the technician’s machine execute it.

RelayTrap replaces the legitimate Guest process on an infected endpoint with a malicious copy. A companion component, TrapWatch, watches ScreenConnect services and compiles RelayTrap from a C# source file using PowerShell, so the attackers do not need to ship a prebuilt executable. When a technician connects, RelayTrap waits about 2.5 seconds and then sends a batch file called Run-Lateral-Full.bat back through the session with the file action set to Run.

The target is CVE-2026-84869, a critical client-side flaw that ConnectWise fixed in ScreenConnect 26.6.5. The bug is a missing-authorization problem: under the right conditions, a Guest can transfer and execute a file through an active session without the Host’s confirmation. ConnectWise gave it a CVSS score of 9.9 and says the servers are not affected. The risk sits on the technician’s workstation.

If the file runs, the next stage tries to steal the technician’s .ASPXAUTH session cookie from the browser. That cookie is proof that the technician has already signed in. With it, an attacker can replay the session against the ScreenConnect panel without a password or a new multi-factor prompt. A component called Panel-Lateral.ps1 does the replay and lists the other sessions the technician can reach. Then ToolboxRelayUploader.exe drops a payload into the panel’s Shared Toolbox, tells every reachable session to run it, and deletes the toolbox item about a minute later.

That is the worm-like design: an infected customer endpoint traps a technician, steals the technician’s management session, and uses the technician’s own permissions to reach other customers. In an MSP, that is the difference between one compromised laptop and a supply-chain incident.

Did the worm actually spread?

Not in the SCKit campaign, as far as the researchers can tell. Blackpoint rebuilt the chain and ran it against a current ScreenConnect client. The malicious Guest queued the transfer, but the Host never requested the file and never started a process. Huntress says the console-takeover attempt failed in every recovered host log it examined.

That is good news, but it is not a reason to relax. The failure is evidence that the patch works against this toolkit, not evidence that the toolkit is harmless. The attackers still installed rogue remote-access clients on hundreds of endpoints. They still had hands-on access. And the vulnerability they tried to use is real.

The flaw is now confirmed exploited

On November 16, Huntress published a separate finding: CVE-2026-84869 has been exploited in the wild. On November 12, the firm saw one MSP technician’s workstation run a Guest-pushed batch file. The file reported to a Telegram bot, downloaded an installer from masterdarky[.]com, and started a rogue ScreenConnect client pointing at the attacker’s relay, all in about nine seconds. A day later, a second technician’s client ran an attacker “canary” file. Huntress also found at least four more technicians exposed in October, including an attempt to drop Sync.bat into a Startup folder through a path-traversal trick.

The scale of the probing is the part that should worry MSPs. Since October 20, Huntress says it has seen at least 688,000 Guest “run file” requests for update*.bat hit 42 MSP ScreenConnect servers from 29 IP addresses. Most were probably automated. The attackers only needed one technician to join the wrong session.

Those incidents are not proof that SCKit’s RelayTrap succeeded. They are proof that the underlying bug is being used, and that a patched client is the difference between a blocked transfer and a compromised workstation.

Why technicians are the target

An MSP technician’s laptop is not just a laptop. It usually has the RMM console, saved credentials, browser sessions, and access to every customer the MSP manages. Stealing that session is often easier than stealing a password, because the technician has already done the hard part: they logged in and passed multi-factor authentication.

That is why the ScreenConnect campaign is a reminder about tier-zero access. The management tools MSPs use to run everyone else’s IT deserve the same protection as domain controllers: strong authentication, least privilege, logging, network isolation, and fast patching. An extension or a support session is part of that trust boundary, not a side feature.

What MSPs should do now

ConnectWise’s November 16 advisory is unusually direct. Its investigation is complete, but partner action is still required. The company says disabling extensions and updating software do not remove access or persistence that is already on a device.

  • Keep the identified extensions disabled: Labtech Plus, RMM Plus, RMM+, RMM Plus Password Link, and RMM+ Password Link. On-premises partners must do this themselves; ConnectWise cannot disable them remotely.
  • Update every technician’s ScreenConnect host client to 26.6.5 or later. Updating the server or cloud console is not enough. Check the installed version on each workstation.
  • Remove the Transfer Files permission from roles that do not need it while any client remains unpatched. ConnectWise lists this as a temporary risk reduction, not a substitute for the update.
  • Rotate keys, API credentials, and service accounts that can reach ScreenConnect. Review extension configuration history and administrator accounts.
  • Hunt for persistence, not just the rogue client: scheduled tasks, registry Run entries, C:\ProgramData\RelayTrap, C:\Windows\Temp\sckit-*.ps1, and Defender exclusions.
  • Invalidate active sessions and force technicians to sign in again. A stolen .ASPXAUTH cookie can survive a password reset.
  • Do not triage a possibly infected endpoint over ScreenConnect. Use a separate security tool, or you risk pulling the trap onto the technician’s machine.
  • Review Run Command and Shared Toolbox history for commands you did not send, and check for unknown Access clients connecting to outside relays.
  • Follow your customer notification obligations. Assume the affected MSP’s customers need to know, even if the vendor has not named them.

The bigger lesson

Remote management tools are force multipliers. That is true for the MSP and for the attacker who takes one over. The industry has seen this pattern before, from Kaseya VSA to the 2024 ScreenConnect authentication bypass. The details change; the economics do not. One management console can reach hundreds of downstream businesses, so it attracts attention.

The SCKit campaign adds a new wrinkle: the malware was designed to attack the person doing the support, not just the device being supported. Even though the worm-like spread failed this time, the design is a warning. The next version may not need a bug. It may only need a technician who trusts the session in front of them.

The bottom line

SCKit is a serious toolkit with a clever idea and, so far, a broken delivery path. The rogue clients and the RMM+ flaw are the parts that actually worked. The technician trap is the part that did not. MSPs should fix both layers anyway: close the extension hole, patch every technician client, and treat support sessions as a possible attack surface rather than a safe administrative channel.

Have a tip? Email [email protected] or use the contact form and choose Editorial. We do not promise anonymity.

Sources: Huntress SCKit analysis; Huntress CVE-2026-84869 update; Blackpoint Cyber; Microsoft Security; ConnectWise advisory; ConnectWise advisory updates.