A shared extension key left empty was enough to drive Run commands across managed fleets. Within about 90 minutes on 11 November 2026, operators on cloud-hosted ScreenConnect panels had pushed rogue clients and the SCKit toolkit onto at least 288 endpoints in 48 organizations, then aimed those implants at technician browser session cookies. According to Huntress, the path that mattered was not a normal ScreenConnect user login. It was the RMM+ extension API, which authenticated with a shared access key and, in versions 1.0.31 and 1.0.32, accepted a blank default when nobody had ever set one.
How blank RMM+ keys opened cloud panels
Huntress tied the burst to RMM+ on ScreenConnect builds 26.6.6.9747 and 26.7.6.9783. The extension’s own setting text called the AccessKey “like a password” and said it had to be sent in the POST body to reach functions. Huntress reported the sharper failure mode in plain language: if an administrator had never configured the extension’s access key, RMM+ would accept a blank one.
That design sat beside a worse architectural choice. According to Huntress, RMM+ bypasses ScreenConnect’s user accounts, roles, and MFA, and authenticates requests with a shared access key stored in the extension’s settings. Panel operators who held that path could issue Run commands that installed rogue ScreenConnect clients and SCKit on managed endpoints. Command activity showed up in logs linked to RMM+, with no matching interactive login for those actions.
ConnectWise’s first public move on the same day did not yet name that flaw. At 22:45 UTC on 11 November 2026, the company reported that a cluster of five third-party LabTech and RMM Plus extensions, including RMM+ and the password-link plugins in that family, had been turned off across roughly 400 partner environments as a precaution, while the initial access vector was still listed as unidentified. A Huntress update around 23:34 ET the same evening connected the activity to the blank default access-key behavior and noted that ConnectWise then turned off RMM+ and related plugins on hosted environments.
The scale was ugly for a single afternoon. Huntress counted at least 288 managed endpoints across 48 organizations. Reporting on the same window put the start near 17:47 UTC and the bulk of impact inside about 90 minutes. Rogue client identifiers reported in the same write-up included 2588ef5c15954b10 and ed630cc2c873b091, with an attacker participant name of “Addmin.” None of that required a published list of victim company names. The MSP blast radius was the point: one cloud panel path, many partner endpoints.
| Control surface | What operators got | Why it failed here |
|---|---|---|
| ScreenConnect user MFA | Never consulted on this path | RMM+ used a shared access key |
| RMM+ AccessKey | Blank default accepted | Versions 1.0.31/1.0.32, empty DefaultValue |
| Run / remote install | Rogue clients plus SCKit | Panel commands without interactive logins |
| Browser .aspxauth cookie | Theft and replay design | Cookie remained portable after sign-in |
What SCKit did on the managed hosts
SCKit was not a random dropper. According to Huntress, the toolkit installed under C:\ProgramData\RelayTrap, registered a TrapWatch scheduled task, and stopped the legitimate ScreenConnect service on the host. Persistence and service disruption were the local half of the job. The remote half was session theft.
SCKit is built to pull the technician’s .aspxauth session cookie from the browser and relay it for replay against the ScreenConnect panel. The intended follow-on was console-side: enumerate sessions, push further payloads, and widen reach across partner environments from a stolen operator session. Huntress is explicit that recovered logs did not show that endpoint-to-technician-workstation console takeover succeeding. Design intent and confirmed spread are not the same fact, and the public record stops at the failed observation for that worm-like hop.
ConnectWise still told partners to treat operator sessions as hostile material. Partners were told to review activity, end active sessions, reset credentials from a device they trust, and confirm multifactor authentication. Huntress telemetry indicated compromised credentials or hijacked operator sessions as a general concern without establishing a patient-zero interactive login method. The company’s disclosure and the Huntress timeline both leave the first human or credential step undetermined. What the telemetry did show was panel access sufficient to Run-install tooling at speed, then implants aimed at cookies that already represented authenticated console use.
Huntress also warned operators not to triage infected hosts over ScreenConnect, precisely to avoid dragging malware onto the technician workstation and outward again. That guidance sits next to the cookie design problem: a browser session cookie that can be copied is a transferable proof of a login that already finished. Ending sessions and rotating credentials are containment after the panel and the endpoints are already in play. They are not a rewrite of the blank-key API path that let Run land in the first place.
Prior rogue-client noise in the same ecosystem
September 2026 already forced Huntress Rapid Response work on worm-like rogue ScreenConnect client activity across several customer environments. ConnectWise later shipped a security update for critical file-transfer issue CVE-2026-84869 (CVSS 9.9). The SCKit reporting cites that wave as historical context in the same ScreenConnect MSP ecosystem and the same class of rogue-client propagation. Public sources do not equate the September activity and the 11 November 2026 SCKit burst as the same attributed actors. They sit as adjacent pressure on the same remote-access fabric, not as a confirmed single campaign identity.
Huntress and ConnectWise have not named a threat actor for the SCKit burst, and individual victim organization names remain unpublished. Public reporting also does not establish interactive login phishing, password theft as patient zero, MFA prompt fatigue, or OTP interception as the panel access path. The confirmed spine is narrower and worse for anyone who assumed console MFA covered extension APIs: a shared key with a blank default, Run-driven rogue clients, SCKit persistence, and a cookie-stealing toolkit pointed at .aspxauth.
Shared API keys that never touch user accounts or MFA are a different failure than a technician typed one-time code on a fake page. Portable console cookies after a legitimate session are a different failure than stopping a login ceremony. Both showed up in the same afternoon’s MSP mess. Hardening the identity story for remote operations still matters when the next path is a real operator login, but it does not rewrite an extension that accepted an empty AccessKey and logged commands under RMM+ with no interactive sign-in beside them.
ScreenConnect RMM+ access keys and session cookie replay prevention.
FAQ
How did attackers reach ScreenConnect panels in the SCKit campaign?
Attackers reached cloud-hosted ScreenConnect panels through the RMM+ extension path that authenticated with a shared access key. According to Huntress, RMM+ versions 1.0.31 and 1.0.32 accepted a blank key when an administrator had never configured one, and that path bypassed ScreenConnect user accounts, roles, and MFA. Operators then used panel Run commands to install rogue ScreenConnect clients and SCKit on managed endpoints.
How many organizations and endpoints did the SCKit ScreenConnect wave hit?
According to Huntress, at least 288 managed endpoints across 48 organizations were affected. Related reporting put the burst starting near 17:47 UTC on 11 November 2026 with roughly 90 minutes of heavy impact. ConnectWise also disabled five third-party LabTech and RMM Plus extensions, including RMM+ and related password-link plugins, across about 400 partner environments as a precaution the same day.
Did SCKit successfully worm from endpoints back to technician consoles?
SCKit is designed to steal the technician browser .aspxauth session cookie and replay it against the ScreenConnect panel, but Huntress reported that recovered logs did not show successful console-takeover spread from endpoint back to the technician workstation. The toolkit still established persistence, stopped legitimate ScreenConnect services, and created RelayTrap components on infected hosts.
Was interactive MFA failure the root cause of the RMM+ access?
No public source establishes login-MFA failure, prompt fatigue, or OTP interception as the RMM+ panel path. Huntress states that RMM+ authenticates with a shared access key in extension settings and bypasses ScreenConnect user accounts, roles, and MFA entirely when that key path is used. ConnectWise and Huntress also left the broader patient-zero method undetermined while documenting the blank default key behavior.
What did ConnectWise tell MSP partners to do after SCKit activity?
ConnectWise turned off the third-party LabTech and RMM Plus extension set that included RMM+ and the password-link variants in hosted partner environments, and told partners to review activity, end active sessions, reset credentials from a device they trust, and confirm multifactor authentication. Huntress separately urged teams not to triage infected hosts over ScreenConnect so malware would not spread onto technician workstations.