ScreenConnect attacks, November 2026: what to do now

Since November 11, 2026, attackers have used MSPs’ ScreenConnect consoles to push rogue ScreenConnect clients and a toolkit called SCKit onto managed computers. Cloud-hosted and on-premises servers are both affected. ConnectWise’s initial assessment pointed to third-party RMM+ family extensions; its November 16 update says its investigation is complete but does not name a confirmed entry route. Responding now? Start with the vendor steps, then run the checks below.

Jump to the steps
By CalMax Systems Security Research · Published November 14, 2026 · Last updated: November 18, 2026 · Checked against ConnectWise, Microsoft, Huntress and Blackpoint (sources below)

What should ScreenConnect admins do now?

If you suspect compromise, isolate affected devices immediately and investigate through EDR or another trusted channel outside ScreenConnect. Include technician workstations that connected to affected endpoints. On-premises admins should disable the identified extensions and update to the latest supported server release. All partners should verify that every technician client runs 26.6.5 or later.

See the vendor’s response steps

Run ScreenConnect on-premises?

Our Secure ScreenConnect playbook is the step-by-step method we used to secure our own on-premises server during this campaign. Do the vendor steps first. The playbook adds to them; it doesn’t replace them.

See the on-premises playbook

Investigation status: checked November 18, 2026. The latest ConnectWise Trust Center advisory update, November 16 at 20:00 UTC, says ConnectWise has completed its investigation and partner action remains required. ConnectWise says completing its investigation does not establish that individual partner or customer environments are unaffected.

What happened in the November 2026 ScreenConnect attacks?

Huntress saw two waves of attackers using ScreenConnect’s Run Command feature to install extra, attacker-controlled clients on managed endpoints. The numbers are what Huntress could see, not the whole campaign. The toolkit also tries to spread to technicians who connect, but researchers have not seen that succeed.

  • November 11, about 17:47 UTC: Huntress saw at least 288 endpoints across 48 organizations affected within about 90 minutes (Huntress). ConnectWise published its first advisory at 22:45 UTC that day.
  • November 12: ConnectWise said it had found about 400 partner environments with at least one RMM+ family extension installed and had disabled the extensions in them where it could. It says that number reflects environments with the extensions installed, not confirmed compromises (ConnectWise advisory).
  • November 13, 16:12 UTC: Huntress detected another wave affecting 205 hosts across 41 organizations, including an organization with a self-hosted instance of the RMM+ plugin (Huntress update). Blackpoint also reported a second wave that appears to target on-premises ScreenConnect instances. ConnectWise’s update that day (21:15 UTC) says researchers are reporting activity against on-premises environments where the extensions are still enabled.
  • November 16, 20:00 UTC: ConnectWise said its investigation is complete and that researchers continue to report activity against on-premises environments where the extensions remain enabled and ScreenConnect has not been updated.
  • Attempted spread: Huntress and Blackpoint have not observed successful endpoint-to-technician propagation. Blackpoint’s test against a current client also did not execute the payload (Blackpoint analysis). Separately, Huntress reported on November 16 that it confirmed in-the-wild exploitation of CVE-2026-84869 on MSP technician workstations, in activity it does not tie to SCKit (Huntress). Huntress also found attacker-controlled guests in legitimate customer support sessions. When scoping exposure, review guest activity during technician connections in those sessions as well as unfamiliar unattended sessions. Clients on 26.6.5 or later have the fix.
  • Extension flaw: Huntress found that RMM+ 1.0.31 and 1.0.32 accepted a blank key when none was configured. ConnectWise’s disclosure describes a static default API secret as CVE-2026-127063. Published November 12 with a CVSS 4.0 score of 9.3, the disclosure concerns the third-party extension and lists no fixed version. Huntress says RMM+ authenticates requests with its own shared access key stored in the extension’s settings and bypasses ScreenConnect’s user accounts, roles and MFA. This flaw is not confirmed as the entry route in every incident.
  • Entry route not confirmed: ConnectWise’s initial assessment pointed to abuse of independent third-party extensions. It found no evidence that its own infrastructure was compromised or that initial access involved a ScreenConnect vulnerability. Its November 16 update says the investigation is complete but does not name a confirmed initial access method (ConnectWise advisory).

What does ConnectWise say to do?

Affected versions and fixes
IssueAffectedFix or mitigation
CVE-2026-84869 (CVSS 3.1 score 9.9)ScreenConnect clients before 26.6.5 in cloud and on-premises deployments. A condition in the client may allow files to be transferred and executed through an active remote session without authorization or Host confirmation. ScreenConnect servers are not affected.26.6.5 or later (NVD build cutoff 26.6.5.9742). Reinstall host clients, update Access agents and verify technician workstations.
CVE-2026-127063 (CVSS 4.0 score 9.3)The CVE record lists all versions of the third-party DataComm RMM+ extension. Deployments retaining the default secret are exposed to unauthorized extension API access.No fixed RMM+ version has been announced. Verify the extension is disabled across your instances and keep it disabled pending vendor guidance.

Huntress observed the November campaign on ScreenConnect builds 26.6.6.9747 and 26.7.6.9783. The client fix addresses the attempted spread to technicians. Extension disablement and incident response address separate risks. See the ScreenConnect vulnerability guide for patch and upgrade details.

From the ConnectWise advisory updates of November 12, 13 and 16. If you suspect compromise, investigate with your EDR or another tool outside ScreenConnect, and include technician workstations that connected to affected endpoints. This list is containment. Scoping an intrusion is a job for your security team or provider.

  1. Disable these extensions and verify their status across every instance: Labtech Plus, RMM Plus, RMM+, RMM Plus Password Link and RMM+ Password Link. ConnectWise disabled them where it could in its hosted environments. On-premises admins must disable them immediately. All partners with these extensions should verify they remain disabled and leave them disabled pending vendor guidance.
  2. Disable the precautionary three too: ConnectWise disabled ITBoost Integration, RepairShopr Integration and Guest Session Starter in its cloud and marketplace, and strongly encourages all partners to do the same until further notice.
  3. On-premises: update ScreenConnect to the latest supported version. In Administration > Overview, check your current and eligible versions. If your license is out of maintenance and does not cover the required release, upgrade the license before installing that release. ConnectWise requires 25.4 or later for a direct upgrade to 26.6.5. If you run an older release, follow the vendor upgrade path. For on-premises ScreenConnect integrated with Automate, ConnectWise provides 26.6.5 through Automate Product Updates for partners with active Automate Assurance.
  4. From a trusted device, revoke ScreenConnect and related identity-provider sessions, then reset credentials. A password reset alone does not end every session.
  5. Enforce MFA for every ScreenConnect user.
  6. Cloud and on-premises: reinstall host clients, update Access agents and verify every technician client is on 26.6.5 or later. Do this after the server update. ConnectWise has already updated its cloud servers. Check the installed version on every technician workstation.
  7. Review Run Command, Toolbox and extension API activity. Investigate unexpected commands, new admin accounts and permission changes. After patching, remove unrecognized users and review existing roles and permissions. If updating must wait, temporarily remove the TransferFiles permission (formerly TransferFilesInSession) under Administration > Security > Roles. Check each assigned session group in every role and save the changes. Apply the update as soon as possible.
  8. Isolate affected devices and preserve evidence before cleanup. After preserving evidence, clear unauthorized queued commands and tools, remove unauthorized software and persistence, and validate remediation before reconnecting devices. If you suspect the ScreenConnect server itself was compromised, isolate it, preserve a backup for analysis, and keep it offline until it has been investigated, rebuilt and patched. If a technician workstation was compromised, rotate credentials that were accessible from it, including RMM, PSA, ScreenConnect, vault and customer administrator credentials.
Containment is not cleanup. Disabling extensions and updating clients do not remove access already established on endpoints. Start with the compromise checks.

What we saw on our own ScreenConnect server

Our fully patched on-premises server had no extensions, yet showed unattended sessions we did not recognize. We do not know whether those sessions were related to SCKit.

How can I secure an on-premises ScreenConnect server?

Do the vendor steps first. After that, hardening the server is its own job, separate from incident response: if you found unauthorized activity, a configuration change does not close the incident. The method we used on our own server is in the Secure ScreenConnect playbook.

Preserve ScreenConnect evidence before cleanup

Save console and endpoint records before you remove anything. Microsoft’s response guidance says to retain extension, authentication, Run Command and Toolbox records together with process, file, task and registry history, and to verify the complete state of each device before restoring connectivity.

Was your ScreenConnect compromised? Quick checks

Check the console first, then the endpoints. Compare what you find with the current indicators and hunting queries in the Microsoft SCKit post and Huntress report. No match does not mean you are clean. The indicators keep changing.

  • Console activity nobody on your team ran: Run Command or Toolbox entries, especially ones that install another ScreenConnect client, plus new admin users or role changes.
  • Installer markers: a c=Exploited marker in installer download commands, such as c=Exploited809 or c=ExploitedNOTUSA.
  • Unattended sessions you can’t match to a real customer machine.
  • An unapproved ScreenConnect client on an endpoint. Validate each client’s enrolled instance and relay destination before removing it. ConnectWise and Microsoft list the rogue client IDs 2588ef5c15954b10 and ed630cc2c873b091.
  • RelayTrap persistence: a RelayTrapLiteAlert Run value, a scheduled task named TrapWatch, a C:\ProgramData\RelayTrap\ folder, or %TEMP%\sckit-* staging folders.
  • New Microsoft Defender path exclusions for Windows Temp, ConnectWiseControl_, CWInstall_ or RelayTrap.

Crypto-named sessions: bait or intruder?

Not every unknown session is an SCKit intrusion. On November 13, Huntress reported unfamiliar sessions named after crypto exchanges, such as SRV-BINANCE, KUCOIN-SRV or BACKUP-PROBIT, on at least 118 self-hosted ScreenConnect servers belonging to 109 organizations it protects. It had not seen malicious activity linked to them and found no overlap with the published SCKit indicators. Its leading hypothesis is that they are bait for CVE-2026-84869, the client flaw fixed in 26.6.5, and it could not rule out other goals.

Huntress says technicians should not join sessions they cannot tie to a customer, and that crypto-exchange names, blank company fields or a mismatched client version should trigger a pause and investigation. Update every technician client to 26.6.5 or later. Don’t connect to an unknown session to take a look. Preserve a copy of App_Data\Session.db and its write-ahead log, then delete confirmed unauthorized sessions from the console. Review the preserved session history to see whether any technician joined and whether files, commands or text were sent. If someone already joined with an unpatched client, investigate that workstation through trusted security tooling and review the preserved server records to determine what occurred during the connection. If someone typed credentials into the session, treat them as potentially exposed and rotate them from a trusted device.

Where an IP blocklist helps

A firewall blocklist can reduce scanning and login attempts from listed addresses. Address the ScreenConnect risks by applying the vendor updates, disabling the identified extensions and investigating suspected compromise.

ThreatListPro covers 5.4M+ IPv4 addresses from real attacks on live production Palo Alto GlobalProtect VPN portals in the US and is checked for new attackers every 6 hours. You can test the free sample first or start a 14-day free trial.

The Secure ScreenConnect playbook

The Secure ScreenConnect playbook is a $79 PDF with implementation steps, testing and rollback guidance. Buyers get free email support. It is for on-premises deployments and does not apply to ScreenConnect Cloud. There are no refunds. Read the product details before buying to check that it fits your setup.

Get the PDF for $79

ScreenConnect attack FAQs

Is there a patch for the November 2026 ScreenConnect attacks?

Partly. Client version 26.6.5, released September 8 to fix CVE-2026-84869, blocks the toolkit's attempted spread from an endpoint to a technician, so update every technician client. The initial access method remains unconfirmed. ConnectWise's November 16 update also tells on-premises partners to update ScreenConnect to the latest supported version. No fixed RMM+ extension version has been announced, so keep the named extensions disabled.

Does this affect on-premise ScreenConnect?

Yes. Early reports were about cloud-hosted consoles, but ConnectWise's November 13 and November 16 updates say researchers are reporting activity against on-premises environments where the named extensions are still enabled. ConnectWise cannot disable extensions on your own server, so you have to do it. Having an extension installed is a reason to act, not proof you were compromised.

Can I protect my on-premises ScreenConnect server?

Yes. Do the vendor steps first: disable the named extensions, update the server to the latest supported version and technician clients to 26.6.5 or later, revoke sessions and enforce MFA. You can then harden the server itself. Our Secure ScreenConnect playbook is the method we used on our own server. Hardening does not clean up an intrusion that already happened.

What is SCKit?

SCKit is the name researchers use for the toolkit in the November 2026 ScreenConnect attacks. It installs an extra ScreenConnect client that connects to attacker infrastructure, sets up persistence, and includes RelayTrap, which targets technicians who connect. That extra client works on its own, so removing a session from your console does not clean the endpoint.

What did attackers breach, and how did they get in?

ConnectWise confirmed unauthorized access to partner ScreenConnect environments. It found no evidence that its own infrastructure was compromised or that a ScreenConnect vulnerability was used for initial access. Attackers did get into MSPs' ScreenConnect consoles and pushed rogue clients from them. ConnectWise's initial assessment pointed to abuse of third-party extensions in the RMM+ family. Its November 16 update says its investigation is complete but does not name a confirmed initial access method.

Are unknown machines in my ScreenConnect session list a hack?

Not necessarily, but treat them as a risk. Huntress found sessions named after crypto exchanges, such as SRV-BINANCE or KUCOIN-SRV, on at least 118 self-hosted ScreenConnect servers. It saw no malicious activity linked to them and no overlap with SCKit indicators, but its leading hypothesis is that they are bait for CVE-2026-84869. Don't join sessions you can't tie to a customer. Commands you didn't run, a second client on a real endpoint or new admin users point to an intrusion.

Is ScreenConnect safe to use right now?

ConnectWise found no evidence that a ScreenConnect vulnerability was used for initial access. It did link SCKit’s attempted spread to the client flaw fixed in 26.6.5. Follow the vendor steps above. If you suspect compromise, isolate affected devices and use EDR or another trusted channel to investigate before restoring connectivity.

How do I remove a rogue ScreenConnect client?

Treat it as malware. Preserve evidence first, then remove the rogue client and its persistence with your EDR or another tool outside ScreenConnect. Uninstalling the client alone is not enough. Microsoft warns not to run cleanup scripts recovered from SCKit, because they can redeploy the toolkit. Check the device before you reconnect it, and rebuild it if you can't confirm it's clean.

Sources

Primary sources reviewed November 18, 2026.

Related resources

General VPN and remote-access reading.

BeyondTrust Remote Support attacks, 2026: what to do now SimpleHelp attacks, 2026: what to do now N-central attacks, 2026: what to do now RDP Brute Force Protection VPN Brute Force Attacks Explained