A blank extension access key drove ScreenConnect Run installs without a single interactive MFA challenge. According to Huntress, the 11 November 2026 SCKit burst reached at least 288 managed endpoints across 48 organizations in roughly 90 minutes after operators used cloud-hosted panels through the RMM+ path. Device-bound phishing-proof login controls would not have closed that shared-key API, and they cannot unsteal a technician .aspxauth cookie after a session already exists. Prevention here is API authorization design, cookie and session hygiene, and clear boundaries on what console MFA actually covers.

Why ScreenConnect user MFA never saw the RMM+ Run path

Huntress observed operators on multiple cloud-hosted ScreenConnect panels pushing Run commands that installed rogue ScreenConnect clients and the SCKit toolkit on managed hosts. Reporting put the start near 17:47 UTC on 11 November 2026, with bulk impact inside about 90 minutes. ScreenConnect builds 26.6.6.9747 and 26.7.6.9783 appeared in the same analysis, alongside rogue client identifiers such as 2588ef5c15954b10 and ed630cc2c873b091 and an attacker participant name of “Addmin.”

ConnectWise’s first public step at 22:45 UTC the same day disabled five third-party LabTech and RMM Plus extensions, including RMM+, Labtech Plus, RMM Plus, and the related password-link plugins, across roughly 400 partner environments while the initial access vector was still listed as unidentified. A Huntress update around 23:34 ET tied the activity to RMM+ versions 1.0.31 and 1.0.32 accepting a blank default access key when an administrator had never configured one. ConnectWise then turned those plugins off on hosted environments. Partners were told to review activity, terminate ongoing sessions, rotate credentials on a trusted device, and confirm multifactor authentication.

RMM+ did not fail a login ceremony. According to Huntress, the extension authenticates requests with a shared access key stored in extension settings and bypasses ScreenConnect user accounts, roles, and MFA when that path is used. The product text for AccessKey even called the value “like a password” and said it had to be sent in the POST body to reach functions. Versions 1.0.31 and 1.0.32 accepted an empty default when nobody had ever set a key.

Commands then reached endpoints with no corresponding interactive login activity and appeared in logs linked to RMM+. Panel operators who held that API surface could install rogue clients and SCKit without presenting a technician password, OTP, push, or hardware factor to ScreenConnect’s normal user stack. Console MFA that only guards human sign-in never saw those POSTs.

SCKit’s job on the host was separate and later. Huntress described install under C:\ProgramData\RelayTrap, a TrapWatch scheduled task, and a stop of the legitimate ScreenConnect service. The toolkit is built to pull the technician browser .aspxauth session cookie and relay it for replay against the panel so operators can enumerate sessions and push further payloads. Recovered logs did not show that endpoint-to-technician-workstation console takeover succeeding.

ConnectWise still treated operator sessions as hostile material after the fact. Ending sessions and rotating credentials contain damage after panel access and endpoint implants already exist. They do not rewrite an extension API that accepted a blank shared key. Huntress also warned teams not to triage infected hosts over ScreenConnect, precisely so malware would not land on the technician workstation and move outward again.

Surface What operators used Why console MFA missed it
RMM+ AccessKey Blank default on 1.0.31/1.0.32 Shared key, not user login
ScreenConnect accounts/roles Never consulted on this path Extension bypassed identity stack
Run / remote install Rogue clients plus SCKit Logged under RMM+, no interactive sign-in
Browser .aspxauth Theft and replay design Cookie already proved a finished session

Huntress and ConnectWise have not published individual victim organization names or a named threat actor for this burst. Public reporting also leaves the broader patient-zero method undetermined. The confirmed spine is panel access sufficient to Run-install tooling at speed, then implants aimed at operator session material. For the minute-by-minute attack narrative, read the companion on how blank RMM+ keys opened ScreenConnect panels.

September 2026 already produced Huntress Rapid Response work on worm-like rogue ScreenConnect client activity in the same MSP ecosystem, and ConnectWise later shipped a fix for critical file-transfer issue CVE-2026-84869 (CVSS 9.9). Sources cite that wave as historical context, not as proven identical actors to the November SCKit burst.

Where device-bound phishing-proof login helps after SCKit

The RMM+ stage stole authorization through an API shared key outside interactive authentication. MFA 2.0 is phishing-proof device-bound authentication built on public-key cryptography: there is no central password database to steal, and signatures stay bound to an enrolled device. A phishing-proof IdP that only issues hardware-bound signatures for human sign-in still never sees a blank AccessKey POST. Closing that gap is extension configuration, removal of default empty secrets, least privilege on automation APIs, and vendor patch or disablement of the broken plugin path. ConnectWise’s disablement across hundreds of partner environments was the emergency version of that control.

The SCKit stage targeted a replayable browser session cookie after authentication had already produced a valid console session. No login MFA undoes cookie theft from an endpoint that already holds a live .aspxauth value. Prevention value for remote phishing kits is upstream at sign-in. Once malware or a toolkit reads a cookie from a machine, defenders need session revoke, short-lived tokens where the product allows them, technician workstation isolation, and the operational rule Huntress stressed: do not live-triage infected hosts through the same remote tool that may pull malware onto the console.

Device-bound credentials still matter for the next real technician login. When a partner rotates credentials on a trusted device and returns through a normal console sign-in, origin-bound signatures with no transferable OTP or push remove the cheap remote harvest of passwords and codes. That is prevention, not detection: MFA 2.0 removes phishable factors so the remote kit never completes a login. It is a different problem than the blank-key API. Prevention-focused device-bound login helps when a phishable login factor is actually in the chain. Here the first chain link was authorization abuse on an extension key.

Portable console cookies remain a blast-radius problem even when login is clean. A copied .aspxauth value is proof of a session that already finished. Binding sessions to device state, forcing re-auth for sensitive Run actions, and aggressively invalidating operator cookies after suspected panel abuse shrink that window. Those controls sit beside identity, not inside a second-factor prompt.

Control focus              | Blank RMM+ AccessKey     | Stolen .aspxauth cookie
---------------------------|--------------------------|---------------------------
Interactive MFA            | Not invoked              | Already satisfied earlier
Shared API secret hygiene  | Primary fix              | Not the cookie problem
Session revoke / reset     | After panel abuse        | Primary containment
Device-bound login crypto  | Next human sign-in only  | Does not unsteal cookies

Passkeys and FIDO2-style WebAuthn put a key pair on the device, keep the private key in hardware, and answer an origin-bound challenge so a fake login host cannot complete the ceremony. That design is origin-bound at authentication. On a generic enterprise IdP that issues only hardware-bound signatures, the user unlocks a local private key; nothing reusable is typed into a proxy page.

That property would matter if SCKit’s patient zero had been a coached technician login. Public sources do not establish that path for this wave. A origin-bound login also does not validate or replace a shared extension AccessKey living in plugin settings. Automation APIs that accept a static secret still need secret management, non-empty defaults, rotation, and monitoring whether or not console users hold passkeys. MFA 2.0 goes further as phishing-proof across the full identity lifecycle when no phishable enrollment or recovery factor remains, yet neither passkeys nor MFA 2.0 replace AccessKey policy on RMM tooling.

This campaign’s public record does not turn on helpdesk re-enrollment, Temporary Access Pass readout, or a phishable device-onboarding code. The identity-adjacent failures were an API key that never touched user enrollment state and a browser cookie minted after a finished session. Phishing-proof coverage across registration, onboarding, authorisation, authentication, and decommissioning still leaves shared automation secrets and post-auth cookie theft as separate engineering problems. The device-bound architecture closes transferable factors in the human identity lifecycle. It is not a substitute for extension AccessKey policy on RMM tooling.

MSP defenders should still harden recovery so the cleanup ConnectWise requested cannot be social-engineered into attacker-controlled factors. That cleanup includes credential rotation on a trusted device and MFA verification. That is good hygiene after panel abuse. It is not a rewrite of versions that accepted a blank key.

Huntress telemetry mentioned compromised credentials or hijacked operator sessions as a general concern without establishing interactive login phishing, password theft as patient zero, prompt fatigue, or OTP interception as the panel path. Those remain open. The documented failure modes that do not need speculation are the empty AccessKey default and the cookie-stealing implant design.

Key takeaways for defenders

Shared keys that skip accounts and MFA fail differently than a typed one-time code. Portable console cookies fail differently than a blocked login ceremony. Both showed up in one MSP afternoon. Build the identity stack so the next human sign-in has no transferable factor, and build the remote-ops stack so blank AccessKeys and replayable .aspxauth values are not standing invitations.

FAQ

Would phishing-proof MFA have stopped the RMM+ blank AccessKey path in the SCKit campaign?

No. Phishing-proof MFA would not have stopped the RMM+ blank AccessKey path in the SCKit ScreenConnect campaign because that path never called ScreenConnect user authentication. According to Huntress, RMM+ authenticated with a shared access key in extension settings, bypassed accounts, roles, and MFA, and in versions 1.0.31 and 1.0.32 accepted a blank default when no key had been configured. Fixing that surface is non-empty secrets, plugin disablement, and API authorization design, not a stronger human login factor.

Can device-bound MFA undo technician .aspxauth cookie theft after SCKit lands?

Device-bound MFA cannot undo technician .aspxauth cookie theft after SCKit lands on a managed host. SCKit is designed to read a browser session cookie that already represents a finished console login and replay it against the ScreenConnect panel. Login cryptography only decides whether a new sign-in completes. Once the cookie exists on an endpoint, containment means revoking sessions, rotating credentials on a trusted device, and isolating technician workstations from infected hosts.

What ScreenConnect extension controls should MSPs prioritize after this wave?

MSPs should prioritize inventory and hardening of every third-party extension that uses a shared access key, starting with the LabTech and RMM Plus family ConnectWise disabled across about 400 partner environments on 11 November 2026. Require strong non-default AccessKey values, remove plugins that are not essential, alert on Run commands logged under extension identities without interactive logins, and keep operator session revoke playbooks ready. Huntress counted at least 288 endpoints across 48 organizations in this burst, which is enough scale to justify treating extension keys as tier-0 secrets.

How do passkeys differ from fixing shared RMM API keys on ScreenConnect?

Passkeys differ from fixing shared RMM API keys because passkeys protect the human authentication ceremony with origin-bound, device-held private keys, while shared RMM API keys are static secrets checked by extension code. Passkeys are origin-bound at login on a hardware-bound enterprise IdP. MFA 2.0 goes further as phishing-proof across the full identity lifecycle when no phishable enrollment or recovery factor remains. Neither replaces AccessKey policy, default-value bans, or monitoring for automation that bypasses user accounts on ScreenConnect.

What did public sources leave open about SCKit initial access?

Public sources left the broader patient-zero method open for the SCKit ScreenConnect campaign. Huntress and ConnectWise documented the blank RMM+ access-key behavior and the cookie-stealing toolkit, but they did not establish interactive login phishing, password theft as the first step, MFA prompt fatigue, or OTP interception as the panel entry. They also did not show successful console-takeover worm spread from endpoint back to the technician workstation in recovered logs, and they did not name a threat actor or individual victim companies.