SCKit Worm Hijacks ScreenConnect to Backdoor MSPs and Their Customers — 493 Endpoints, Two Waves, No CVE
A custom malware toolkit called SCKit hijacked ScreenConnect remote management panels to push persistent backdoors onto 493 managed endpoints across 89 organizations in two waves. The root cause was a blank authentication key accepted by a third-party extension. This is a supply-chain attack on the MSP layer itself — your organization may be a victim even if you never touched ScreenConnect.
What: SCKit, a 16-file custom malicious toolkit, deployed through compromised ScreenConnect remote management panels via abuse of a third-party RMM+ extension with a blank-access-key authentication flaw. No CVE has been assigned. This is an active campaign, not a vulnerability waiting to be exploited — it is being exploited right now.
Status: Active. Wave 1 hit November 11, 2026. Wave 2 hit November 13. ConnectWise has disabled five third-party extensions across 400+ partner environments. Huntress reports ongoing monitoring. No end-of-campaign declaration has been issued.
Deadline: No CISA KEV entry and no federal agency deadline. The operational deadline is today: if you or your MSP runs ScreenConnect and has not audited Run Command history, you may have an active backdoor on your endpoints right now.
Fix: Disable RMM+, Labtech Plus, RMM Plus, RMM Plus Password Link, and RMM+ Password Link across all ScreenConnect instances. Audit Run Command and Shared Toolbox history. Check for a second unauthorized ScreenConnect client and the TrapWatch scheduled task. Isolate suspected hosts and investigate with out-of-band tools — not through ScreenConnect.
Key takeaways
- Your MSP is the attack surface. SCKit did not need to compromise your organization directly. It compromised the managed service provider whose ScreenConnect panel reaches your endpoints. If you use an MSP that runs ScreenConnect, ask them today whether they were affected.
- 493 endpoints hit across 89 organizations in two waves. Wave 1 (November 11) hit 288 hosts across 48 organizations in roughly 90 minutes — about a third in the first 10 minutes. Wave 2 (November 13) added 205 hosts across 41 organizations with new infrastructure (single-source: Huntress).
- The entry point was a blank authentication key. RMM+ extension versions 1.0.31 and 1.0.32 defaulted to an empty
RMMPlus.AccessKeyand accepted an empty value as valid, bypassing ScreenConnect user accounts, roles, and MFA entirely — leaving no login events in audit logs (single-source: Huntress). - The toolkit appears partially LLM-generated. Huntress analysts noted that
RelayTrap.csshows structural patterns and inline comments consistent with AI-assisted code generation. The attackers are using AI tooling to build custom malware faster. - Lateral movement was designed but not confirmed. The toolkit included code designed to steal technician session cookies and spread to other panels, but Huntress found no confirmed case where this portion succeeded in recovered host logs.
- Removal requires out-of-band access. Do not triage or investigate over ScreenConnect. Use WinRM, EDR live response, or another independent channel. Treat any technician workstation that connected to a suspected host as potentially exposed.
Executive Summary
The MSP supply chain is under active attack — and the entry point was a default configuration that nobody noticed was unlocked.
Huntress researchers detected suspicious ScreenConnect activity on November 11, 2026, beginning around 17:47 UTC. Within 90 minutes, 288 managed endpoints across 48 customer organizations had been hit. The attack used ScreenConnect’s own Run Command feature to silently install a custom malicious toolkit — SCKit — that planted a persistent second ScreenConnect client under attacker control, disabled Defender’s visibility, and built a “relay trap” designed to intercept session cookies from technicians who later connected to the infected machines.
ConnectWise, which makes ScreenConnect, confirmed on November 12 that five third-party extensions were likely involved and disabled them across more than 400 partner environments. It found no evidence of a vulnerability in ScreenConnect itself, and no CVE has been assigned. The root cause, identified by Huntress in the RMM+ extension, was an authentication check that accepted a blank passphrase as valid — a default that turned an operator-level API into an open door.
A second wave on November 13 hit 205 more hosts across 41 organizations using new infrastructure. As of November 15, the campaign has not been declared over.
The question for every enterprise today is not whether ScreenConnect has a patch to install. The question is whether your MSP’s panel was one of the compromised ones — and whether you have an unauthorized remote-access client running on your endpoints right now.
Why This Matters to CISOs and Security Leaders
This is not a software vulnerability story. It is a supply-chain trust story.
When you hire a managed service provider, you implicitly trust every tool they use to manage your environment. ScreenConnect is one of the most widely deployed remote management platforms in the MSP market. The 48 organizations hit in wave 1 almost certainly had no idea their MSP was running an extension with a blank password protecting an administrative API. They had no CVE to watch, no patch to apply, no alert to investigate. They were simply downstream of someone else’s configuration failure.
The scale multiplier is what makes MSP-layer attacks so dangerous. Huntress reported that one single compromised ScreenConnect panel reached 190 downstream client environments (single-source: Huntress). A single operator credential or a single blank passphrase in one MSP shop can touch hundreds of small and mid-size enterprises simultaneously — the same economics that made the N-able N-central zero-day two days earlier a critical MSP-layer risk.
The LLM-generated attack code is a second signal worth flagging to boards. The toolkit’s C# component shows structural and commentary patterns that Huntress analysts associate with AI-assisted development. This aligns with a broader trend — Aur0ra ransomware’s AI agents reaching domain admin was covered here on November 3. Threat actors are using AI tooling to build and iterate on custom malware faster than defenders can track it.
Ask your MSP two questions before the end of today: “Do you use ScreenConnect, and did you disable the five listed third-party extensions before November 11?” If the answer to the first is yes and the answer to the second is no, treat your environment as potentially compromised and start an investigation, not a patching exercise.
The Timeline
| Date / Time (UTC) | What happened |
|---|---|
| September 8, 2026 | ConnectWise releases ScreenConnect 26.6.5, fixing CVE-2026-84869 (CVSS 9.9), a prior file-transfer flaw. Huntress had already responded to similar activity in an earlier wave. |
| Nov 11, ~17:47 | SCKit wave 1 begins. All 288 endpoints across 48 organizations are hit within approximately 90 minutes; roughly a third within the first 10 minutes. |
| Nov 11, 22:45 | ConnectWise advises that five third-party extensions may be involved and begins disabling them across 400+ partner environments. |
| Nov 12 | Huntress publishes a full technical report. Sophos reports unauthorized deployments from three specific ScreenConnect instance IDs. ConnectWise says no ScreenConnect vulnerability has been found. |
| Nov 13, ~16:12 | SCKit wave 2 begins. 205 hosts across 41 organizations hit with new attacker infrastructure, including a port-proxy redirect and hosts-file modification (single-source: Huntress). |
| Nov 13–14 | Multiple security media outlets publish coverage. HHS issues a sector alert for healthcare organizations using MSPs. No threat actor attribution published. |
| Nov 15 (today) | Campaign ongoing. No end-of-campaign declaration. Combined scope: at minimum 493 endpoints across 89 organizations across two confirmed waves. |
The attackers didn’t need a zero-day. They needed a blank password that everyone forgot to set.
Attack Chain Analysis
The following breakdown is drawn from Huntress’s primary technical report and confirmed independently by ConnectWise’s advisory and reporting from The Hacker News and eSecurity Planet.
| Stage | Technique | Defensive control that matters |
|---|---|---|
| Initial access | Attacker uses RMM+ extension API, authenticated with a blank passphrase accepted by default in versions 1.0.31 and 1.0.32. Bypasses ScreenConnect user accounts, roles, and MFA. No login events appear in audit logs (single-source: Huntress). | Disable RMM+ and related extensions. If keeping on-premises RMM+, set a strong, non-blank access key. |
| Execution | Operator triggers ScreenConnect’s Run Command feature to run commands as SYSTEM on all reachable endpoints. Command installs an attacker-controlled sc.msi from C:\Windows\TEMP\ and drops sckit-run.bat. |
Alert on mass Run Command activity from a single panel session. SRP/AppLocker rules blocking temp-folder execution will catch this. |
| Payload delivery | sckit-run.bat downloads sckit-decoder.ps1 and a machine-specific encrypted SCKit package from attacker infrastructure. The server returns a decryption key unique to each machine UUID. |
Block outbound connections to unknown hosts from managed endpoints. EDR detections on PowerShell base64 decode + network fetch chains. |
| Persistence | SCKit extracts to C:\ProgramData\RelayTrap\. A second attacker-controlled ScreenConnect client is installed. TrapWatch.ps1 runs every 10 seconds as SYSTEM, compiles a host-specific RelayTrap.exe, creates a TrapWatch scheduled task, stops the legitimate ScreenConnect service, and sets it to manual start, substituting the attacker’s copy. Defender exclusions are added. Wave 2 also added a netsh port-proxy redirect and a hosts-file modification (single-source: Huntress). |
Alert on new scheduled tasks named TrapWatch, new services named “Windows Update Agent,” Defender exclusion changes, and ScreenConnect service transitions to manual. |
| Designed lateral movement (not confirmed in incidents) | When a technician connects to an infected endpoint, the toolkit attempts to steal the technician’s browser session cookies (.aspxauth), use them to log into the ScreenConnect panel, upload a payload to the Shared Toolbox, and broadcast it to all reachable sessions before deleting the entry ~60 seconds later. Huntress found this design in the code but confirmed it failed in every recovered host log it examined. |
Revoke all ScreenConnect web sessions. Rotate credentials from a clean device. Isolate infected endpoints from the panel before allowing technician access. |
| Command and control | Attacker-controlled ScreenConnect clients connect to panels hosted on bulletproof infrastructure (1337 Services GmbH). Wave 2 added port-proxy redirection to 185.241.211[.]197:80. |
Block the IoCs listed below. Monitor for outbound ScreenConnect traffic to unauthorized panel domains. |
The wave 2 infrastructure shift happened within 48 hours of wave 1 becoming public. The attackers deployed new IPs, added a hosts-file modification, and switched up their port-proxy approach. This is not a one-and-done campaign — it is an adversary actively adapting. IoCs published on November 12 may already be stale for wave 2 victims. The more durable detections are behavioral: TrapWatch task creation, unauthorized ScreenConnect service state changes, and mass Run Command events from a single session.
Threat Intelligence Breakdown
Attribution: Unknown. No named threat group has been identified. Huntress notes the attacker’s ScreenConnect participant name was “Addmin” (sic), and the infrastructure traced to 1337 Services GmbH, a bulletproof hosting provider. The LLM-generated code in RelayTrap.cs does not indicate a specific actor but confirms capability investment. The sequential naming of MSP panels (“Exploited1,” “Exploited2,” etc.) reported by The Hacker News suggests the attacker had been targeting MSPs methodically before the detected waves.
What is still unknown: How the attacker initially discovered the blank-key RMM+ installations. Whether wave 1 and wave 2 are operated by the same actor or different actors. Whether there are earlier, undiscovered waves. The total number of organizations that may have had the vulnerable extension installed.
Indicators of compromise (Huntress primary report, defanged; wave 2 IoCs single-source):
| Type | Indicator | Wave / Context |
|---|---|---|
| IP | 185.99.255[.]28 | Wave 1 — attacker ScreenConnect server panel |
| IP | 155.94.150[.]39 | Wave 1 — attacker ScreenConnect panel (sc.payproofkit[.]com) |
| IP | 185.241.211[.]197 | Wave 2 — hosts rogue ScreenConnect client; 1337 Services GmbH |
| Domain | newnewnew[.]gotdns[.]org | Wave 1 — relay resolving to 185.99.255[.]28 |
| Domain | sc.payproofkit[.]com | Wave 1 — attacker panel domain |
| Domain | soxnewwork[.]homedns[.]org | Wave 1 — relay on port 81 |
| Domain | 007[.]from-ak[.]com:88 | Wave 2 — relay domain (single-source) |
| Domain | cdn007-digitaldownloads[.]com | Wave 1 — C2, originally flagged by BlackPoint |
| URL | hxxps://ns.mo7sen[.]dev/EadF.php?p=1 | Wave 1 — Stage0 payload stager |
| URL | hxxps://ns.mo7sen[.]dev/EadF.php?p=4 | Wave 1 — decoder stager |
| File path | C:\ProgramData\RelayTrap\ |
Both — SCKit extraction directory |
| Scheduled task | TrapWatch |
Both — persistence mechanism |
| Service name | Windows Update Agent |
Wave 1 — rogue ScreenConnect disguise |
| Registry | RelayTrapLiteAlert Run key |
Wave 1 — additional persistence |
| Hash (SHA-256) | d2ed7a7bf4f8ac9c7969e9b3a07ebd4dda0d858f372c876ac0a041611698de9a |
TrapWatch.ps1 |
| Hash (SHA-256) | 8cb5f48bbe70b5248db7f327f3d1ad6649df5c767d2a7e0711ee147fd92e06af |
ToolboxRelayUploader.exe |
| Hash (SHA-256) | 7a9aca9fdd4dcc13fe497cabdb13ca1e5d6a529271c8aa495e5ca1c1c80271ec |
Panel-Lateral.ps1 — lateral movement component |
| Hash (SHA-256) | 3c7a98f0fdb6235b8076319c74dc1085f7b349b6d750416dd0cadb07d7972ec4 |
RelayTrap.cs — LLM-generated C# source |
| ScreenConnect client IDs | 2588ef5c15954b10, ed630cc2c873b091 |
Both — attacker client identifiers |
| Installer pattern | /Bin/ScreenConnect.ClientSetup.msi?…&c=Exploited<N> |
Wave 1 — sequential MSP numbering |
Note on hash utility: Huntress warns that most SCKit scripts are recompiled or regenerated per victim machine, making file hashes weak long-term indicators. Prioritize behavioral detections over static hash matching.
Enterprise Impact
Even if you have never heard of ScreenConnect, you may be affected. SMBs and mid-market organizations that outsource IT to a managed service provider often have no visibility into what tools the MSP uses. If that MSP ran ScreenConnect with one of the five affected extensions, SCKit may have been installed on your endpoints — running as SYSTEM, with Defender exclusions in place and a fake “Windows Update Agent” service maintaining persistence.
The confirmed impact of a successful SCKit installation includes: a persistent attacker-controlled remote-access channel on each infected endpoint, the ability to execute arbitrary commands as SYSTEM, Defender exclusions that suppress visibility into further activity, and — for wave 2 victims — a hosts-file modification that could redirect traffic or poison resolution for endpoints on that network. The lateral-movement component was designed to escalate through technician sessions to other MSP customer environments, though Huntress did not confirm this phase executing successfully.
The broader supply-chain risk pattern here is consistent with how nation-state and criminal groups have operated for years: compromise a trusted intermediary that already has keys to many kingdoms, and you get volume that no direct-attack campaign can match.
Healthcare organizations should note that HHS issued a sector-specific alert following the discovery. Any regulated data that passed through or was accessible from affected endpoints may have been accessible to the attacker during the period of compromise.
What Security Leaders Should Do Next
Today
- Ask your MSP immediately whether they use ScreenConnect, whether they ran any of the five affected extensions (Labtech Plus, RMM Plus, RMM+, RMM Plus Password Link, RMM+ Password Link), and whether they have checked their Run Command and Toolbox history for unauthorized mass actions since November 11.
- If you operate ScreenConnect yourself, disable all five named extensions immediately. For on-premises RMM+ installations you intend to keep, set a strong, non-blank value for
RMMPlus.AccessKey. - Audit all endpoints for a second unauthorized ScreenConnect client, the
TrapWatchscheduled task, theC:\ProgramData\RelayTrap\directory, theRelayTrapLiteAlertRun registry value, and the service named “Windows Update Agent” running fromC:\ProgramData\Windows Update Agent (…)\. - Block the IoCs listed above at your perimeter and in EDR where hash-based blocking applies. Apply behavioral detections for TrapWatch task creation and ScreenConnect service state changes.
Within 48 hours
- Isolate any endpoint with matching indicators before attempting remediation. Use WinRM or EDR live response — not ScreenConnect — to investigate. Treat any technician workstation that connected to a suspected host as potentially compromised; investigate it from a clean machine.
- Revoke all ScreenConnect web sessions and rotate credentials from a clean device. Revoke browser and identity-provider sessions, since the toolkit was designed to steal session tokens.
- Remediate thoroughly: removal of the rogue client and the TrapWatch task is not sufficient. Remove the entire
C:\ProgramData\RelayTrap\directory, the Run registry value, all temp-folder SCKit files (sckit-*), andsqlite3.exeif present. Verify the legitimate ScreenConnect client is intact. Huntress reports reinfection after partial remediation, so validate fully before returning to service. - Review ScreenConnect instance audit logs for mass Run Command activity, repeated login failures before the event, unexpected extensions, and new admin accounts or permission changes around the November 11 window.
Long-term
- Add MSP tool stacks to your third-party risk program. Require MSPs to disclose the remote management tools, extensions, and authentication configurations they use on your behalf. A blank passphrase on an administrative API is a configuration failure — but you would only know to ask about it if it’s in scope.
- Alert on mass remote management actions. A single ScreenConnect session executing commands on dozens of endpoints in minutes should fire a detection, regardless of whether the session belongs to a known MSP operator.
- Establish out-of-band investigation channels. When an MSP’s remote management tool is itself the attack vector, you need an independent way to reach your endpoints. EDR with live-response capability and a network management path that does not go through the MSP’s tooling are both worth having.
- Monitor for AI-generated malware indicators. Security teams are beginning to build detections that flag suspiciously well-commented or structurally uniform code that may indicate LLM assistance. This is an emerging capability worth tracking.
Stay ahead of the next supply-chain campaign
DataWater’s executive threat briefing surfaces the story your CISO needs to act on — before the board asks about it.
Winners and Losers
| Better positioned | More exposed |
|---|---|
| MSPs that disabled the five named extensions before November 11, or never ran them | MSPs running RMM+ 1.0.31 or 1.0.32 with a blank access key, especially if they haven’t yet audited their Run Command history |
| Enterprises with EDR on MSP-managed endpoints and visibility into scheduled task creation | SMBs and mid-market firms that rely entirely on their MSP’s tooling for endpoint visibility — if the MSP’s tool is compromised, they have no independent view |
| Organizations that blocked the wave 1 IoCs rapidly and have behavioral detections on TrapWatch task creation | Organizations using static hash-based blocking only — the per-machine compilation of SCKit components means hashes are largely unreliable |
| Teams that require their MSPs to disclose extension configurations and conduct third-party risk reviews | Teams that treat MSP access as fully trusted without reviewing what tools and configurations underpin that access |
| ConnectWise, which responded quickly and disabled extensions across 400+ environments | Third-party extension developers who allowed a blank-passphrase default to reach production in a product with SYSTEM-level command execution capability |
Final Executive Takeaway
The ScreenConnect attack is not a patching problem. No CVE exists. No patch is coming. The fix is a configuration audit and a supply-chain conversation.
SCKit hit 493 managed endpoints across 89 organizations in four days, and the campaign appears to still be running. The entry point was a default blank password on a third-party extension that nobody on the attack surface knew to change. The persistence mechanism runs as SYSTEM with Defender excluded, using fake Windows service names, and was designed to spread through every MSP customer environment a technician’s workstation could reach.
Call your MSP today. Audit your endpoints today. The wave 2 infrastructure shift tells you the adversary is not done.
Frequently Asked Questions
What is SCKit and how does it work?
SCKit is a custom malicious toolkit — 16 files including PowerShell scripts, batch files, and a compiled C# component — deployed through compromised ScreenConnect remote management panels. Once an attacker gains access to a ScreenConnect panel via a third-party extension with a blank authentication key, SCKit installs a second unauthorized ScreenConnect client under attacker control, sets up a scheduled task called TrapWatch that runs every 10 seconds as SYSTEM, adds Defender exclusions, and maintains persistence using fake Windows service names. The toolkit also contains code designed to steal technician session tokens and spread to other MSP customers through the panel’s Shared Toolbox, though Huntress did not confirm this lateral-movement phase executing in any recovered incident.
Does SCKit have a CVE number, and do I need to apply a patch?
No CVE has been assigned to the SCKit ScreenConnect campaign as of November 15, 2026. ConnectWise confirmed no vulnerability exists in the ScreenConnect software itself. The root cause was a configuration flaw in a third-party extension called RMM+, which defaulted to a blank access key that the authentication check accepted as valid. There is no patch to apply. The remediation is disabling the five named third-party extensions (Labtech Plus, RMM Plus, RMM+, RMM Plus Password Link, RMM+ Password Link), auditing Run Command history, and removing SCKit components from any affected endpoints.
How do I know if my organization’s endpoints were affected by SCKit?
Search endpoints for the following indicators: a second unauthorized ScreenConnect client not installed by your IT team; a scheduled task named TrapWatch; the directory C:\ProgramData\RelayTrap\; a Windows service named “Windows Update Agent” running from C:\ProgramData\Windows Update Agent (…)\; a Run registry value named RelayTrapLiteAlert; and files beginning with sckit- in the Windows Temp directory. Also review your ScreenConnect Run Command history for mass execution events affecting many endpoints simultaneously, particularly around November 11 and November 13, 2026.
What is the RMM+ extension and why was it dangerous?
RMM+ is a third-party extension for ScreenConnect that exposes a remote-command API for managed service providers. In versions 1.0.31 and 1.0.32, the extension’s access key setting — RMMPlus.AccessKey — defaulted to a blank value, and the authentication check accepted a blank value as valid. Because RMM+ bypasses ScreenConnect’s normal user accounts, roles, and MFA, an attacker who discovered the API endpoint could execute commands across all endpoints managed by that panel without any login event appearing in audit logs. This made the compromise invisible to routine log review. ConnectWise disabled RMM+ and four related extensions across 400+ partner environments on November 11 after the attack was detected. Organizations running ScreenConnect on-premises must disable these extensions manually.
Is the SCKit lateral-movement capability a confirmed risk?
The SCKit toolkit included code designed to steal a technician’s browser session cookie when they connected to an infected endpoint and use it to log into the ScreenConnect panel, upload a payload to the Shared Toolbox, and push it to all sessions reachable from that panel. However, Huntress reported that in every recovered host log it examined, this lateral-movement phase failed to execute. The designed capability is a serious risk for the future and for environments that have not yet cleaned infected endpoints — a technician connecting to an un-remediated host could still trigger the attempt. Revoke all ScreenConnect web sessions and investigate technician workstations from clean devices before allowing reconnection.
What should I ask my managed service provider about this incident?
Ask your MSP four specific questions: First, do you use ScreenConnect to manage our environment? Second, were you running any of the five third-party extensions named by ConnectWise — Labtech Plus, RMM Plus, RMM+, RMM Plus Password Link, or RMM+ Password Link? Third, have you audited your ScreenConnect Run Command and Shared Toolbox history for the period around November 11 and November 13, 2026? Fourth, have you found any unauthorized ScreenConnect clients or the TrapWatch scheduled task on any endpoints you manage for us? A reputable MSP should be able to answer all four questions promptly. Inability to answer suggests their investigation is incomplete.
Why does DataWater say hash-based blocking is unreliable for SCKit?
Huntress noted that the core SCKit component, RelayTrap.cs, is compiled per victim machine — the C# source code is compiled fresh on each infected endpoint, producing a unique binary with a unique hash each time. This means file hash signatures for RelayTrap.exe will differ across every victim. The same applies to other scripts that embed machine-specific values. Static hash-based blocking will catch known hashes from published reports but will not catch variants on new victims. Behavioral detections — such as alerts on TrapWatch task creation, unauthorized ScreenConnect service state changes, Defender exclusion modifications, and mass Run Command events from a single session — are more reliable and durable than hash lists.
Sources
Primary sources: Huntress, “SCKit Worm Installs Rogue ScreenConnect Clients and Session Cookie Relays Across Hundreds of Endpoints” (November 12, 2026 — primary technical source for attack chain, IoCs, timeline, and RMM+ authentication flaw; single-source items noted where applicable). Corroborating coverage: The Hacker News (scope, ConnectWise response, LLM-generated code observation); eSecurity Planet (scope, ConnectWise advisory); IT Security Guru (timeline, speed of infection). Wave 2 details (205 hosts, November 13) are single-source from Huntress. Reflects public reporting as of November 15, 2026.





