This edition focuses mainly on news that directly affects MSPs. Barracuda turns out to have been hacked in early November via a forgotten cloud key, which also put the Barracuda XDR integrations with customers’ Microsoft 365 environments at risk. The ScreenConnect flaw in the technician client is now confirmed as being exploited, and ConnectWise has disabled extensions at 400 partners. In addition, a critical security update for TYPO3 is released today, attackers are exploiting last week’s Microsoft 365 outage for phishing, and Copeland refrigeration controllers for which no patch will be released are under attack.
- Barracuda hacked via forgotten cloud key: XDR customers must recreate Microsoft 365 integrations
- ScreenConnect flaw in the technician client: now actively exploited, ConnectWise disables RMM+ extensions
- TYPO3 releases critical security update for all supported versions today
- Phishing campaign exploits Microsoft 365 outage with fake notifications about held mail
- Copeland refrigeration controllers attacked within two hours of disclosure of 14 flaws, no patch coming
Barracuda hacked via forgotten cloud key: XDR customers must recreate Microsoft 365 integrations
| Vulnerability | Intrusion into Barracuda’s internal cloud environment via a valid cloud access key that Barracuda had given to an external AI vendor and never revoked. One of the affected systems contained the integration credentials that Barracuda Managed XDR Cloud and Monitored Endpoint use to retrieve data from customer environments such as Microsoft 365, Google Workspace and endpoint platforms. No software bug is involved |
| CVE number | No CVE assigned |
| Actively exploited | Yes, intrusion between 1 November 04:33 UTC and 2 November 23:31 UTC, discovered on 4 November. Barracuda has no direct evidence that the integration credentials were accessed, but cannot rule it out either |
| Patch available | No, this is not a software vulnerability. Barracuda has rotated all reachable credentials and contacted affected customers directly; they must revoke and recreate their integration credentials |
How has this been exploited?
Barracuda describes the attack itself in a detailed blog post. The key had been given to an AI vendor for an earlier project and was never revoked afterwards. ‘That is on us’, the company writes. The vendor’s development environment was hacked in October, and according to Dutch security news site Security.nl the key was stored in a configuration file in that company’s GitHub repository. From 1 November onwards, the attacker used every secret found as a stepping stone to the next system: AWS keys and Azure service principal secrets, a data analytics system, a source code token and ultimately Kubernetes clusters. Using short-lived branches, the attacker had build pipelines write secrets to the job logs, encoded to bypass masking. In around 44 hours, some ten cloud accounts were hit and a backdoor container was placed in a Kubernetes cluster. Barracuda links the method to the AI attack framework ARTEX, but does not name a perpetrator. According to SecurityWeek, the attacker also searched for crypto wallets and seed phrases.
What could the consequences be?
The Barracuda XDR integration credentials provide read access to logging and data in customer environments. Because Barracuda’s logging was not detailed enough to rule out access, the company is treating all reachable credentials as potentially compromised. For MSPs deploying Barracuda XDR at customers, this means work on the integrations with Microsoft 365, Google Workspace and EDR platforms. The incident also shows how a key left behind at a third party can affect an entire supply chain.
What should IT administrators do now?
If Barracuda has contacted you, revoke and recreate the integration credentials following the steps in the advisory in the Trust Center. Then check the sign-ins of the applications concerned in Entra ID and Google Workspace for activity from unknown IP addresses since 1 November. If you have not been contacted, Barracuda says no action is required. Do use the opportunity to check which API keys and app secrets you yourself have ever given to vendors or project partners, and revoke anything that is no longer needed.
ScreenConnect flaw in the technician client: now actively exploited, ConnectWise disables RMM+ extensions
| Vulnerability | Flaw in the technician’s host client in ConnectWise ScreenConnect that allows a guest in an active session to send files to the technician and have them executed without authorisation or confirmation. Clients older than 26.6.5 are affected; weeks after the fix, Huntress still saw clients 22.x, 24.2 and 25.8 in use. In addition, an authentication flaw in the third-party extension RMM+ (versions 1.0.31 and 1.0.32) |
| CVE number | CVE-2026-84869 (technician client); no CVE assigned for RMM+ |
| CVSS score | 9.9 |
| Actively exploited | Yes, on 16 November Huntress reported a fully compromised technician workstation (12 November) and 688,000 ‘run file’ requests on 42 MSP ScreenConnect servers between 20 October and 16 November |
| Patch available | Yes, technician client 26.6.5 or later; updating only the server is not enough. Interim measure: revoke the TransferFiles permission. ConnectWise has disabled five extensions in 400 partner environments; on-premises partners must do this themselves |
How has this been exploited?
In the Saturday 14 November edition, the fake sessions bearing crypto exchange names were still suspected bait for technicians with an unpatched client, with no proven malicious activity. Huntress (John Hammond) now confirms real exploitation. Attackers open guest sessions on ScreenConnect servers and continuously send batch files such as update.bat to the server, waiting for a technician to connect. On 12 November, a technician’s client executed such a file, after which the workstation was taken over in about nine seconds: the script reported in via Telegram, downloaded an MSI from masterdarky[.]com and installed a renamed ScreenConnect client that connected to the attacker’s relay. In October, attackers also attempted to use path traversal to place a Sync.bat in technicians’ Startup folders. According to The Hacker News, ConnectWise has now disabled the extensions Labtech Plus, RMM Plus, RMM+, RMM Plus Password Link and RMM+ Password Link in 400 partner environments. On 13 November, a new SCKit wave hit 205 hosts at 41 organisations.
What could the consequences be?
The flaw reverses the usual direction: an infected customer device can take over the technician’s workstation. From that workstation, the RMM, PSA, password vault and admin accounts of other customers are within reach. ConnectWise stresses that updating and disabling extensions do not remove existing access and persistence.
What should IT administrators do now?
Update the ScreenConnect client on every technician workstation to 26.6.5 or later, and if that is not immediately possible, remove the TransferFiles permission from roles and session groups. As an on-premises partner, disable the five extensions listed and do not re-enable them until ConnectWise advises you to. Search a copy of App_Data\Session.db for RanFiles and SentFiles events from guest connections, particularly repeated update*.bat requests, marker.txt, Run-Lateral.bat and paths containing ../. Check technician workstations for wclient_<8 hex> folders, unexpected Startup items and a ‘Windows Update Agent’ service running UpdateConnect.ClientService.exe. If you find anything, isolate the workstation and replace all credentials that were accessible from it.
TYPO3 releases critical security update for all supported versions today
| Vulnerability | Multiple vulnerabilities in TYPO3 CMS Core, the most severe of which is rated critical and can be exploited by backend users with low privileges. All supported versions are affected, including ELTS |
| CVE number | No CVEs disclosed yet; these will follow with the release |
| Actively exploited | No, not reported |
| Patch available | From today: the security releases will be published on Tuesday 17 November at around 12:00 Dutch time (11:00 UTC) |
How can the vulnerability be exploited?
The TYPO3 Security Team pre-announced the release in TYPO3-PSA-2026-002. Details will only follow upon publication, but the team says the most severe flaw can be exploited by users who can already log in to the backend with limited privileges, without any further interaction. The CVSS vector provided points to full impact on confidentiality, integrity and availability, extending beyond the vulnerable component itself. In addition to this flaw, vulnerabilities rated high and medium are being fixed. The German CERT-Bund has listed the issue as critical and unpatched.
What could the consequences be?
In the Netherlands, TYPO3 is widely used by government bodies, educational institutions and larger organisations, often with dozens of editors and external agencies holding a backend account. Using this flaw, a stolen or misused account with limited privileges could escalate to full control of the site and possibly the underlying server. As soon as the patches are public, attackers can compare them with the previous version to work out the flaw.
What should IT administrators do now?
Set aside time today at around 12:00 to update the TYPO3 installations you manage, and read the full advisories on typo3.org at that point for the exact version numbers. Map out in advance which customers run TYPO3 and on which version, so you can roll out immediately. In the meantime, check which backend accounts are still needed and disable old accounts belonging to former employees and agencies. If a customer is running an unsupported version, it will not receive an update and an upgrade is required.
Phishing campaign exploits Microsoft 365 outage with fake notifications about held mail
| Vulnerability | Large-scale phishing campaign using fake Microsoft 365 service health notifications that refer to the genuine Exchange Online outage EX1550543 of 12 November. The emails claim that messages have been held and will be deleted within 24 hours if the recipient takes no action |
| CVE number | No CVE assigned |
| Actively exploited | Yes, emails sent on 13 November; Microsoft Threat Intelligence issued a warning on 16 November |
| Patch available | No, this is not a software vulnerability. Defender for Office 365 blocks and quarantines the emails and Defender blocks the associated domains |
How has this been exploited?
According to Microsoft Threat Intelligence, on 13 November, a day after the genuine outage, attackers sent emails with subject lines such as ‘Outage’, ‘Strengthen’ or ‘Restore’ followed by a random code. The sender name consists of three letters resembling an organisation plus ‘Outage Notification’ or ‘Service Health’. The email contains a hidden fake conversation to mislead filters, and buttons such as ‘RESTORE MESSAGES’ lead via tracking links from services including SendGrid to reallifeaccounting.com, outagefixed.sbs or outagerestored.sbs. The emails came from the domains moments-notice.com, planetmaison.com and newbooker.com. Microsoft assumes the aim is credential theft, although this has not yet been confirmed. Security.nl and SecurityWeek report that the outage itself was not an attack.
What could the consequences be?
The outage was real and many users experienced it, which makes the emails credible. Clicking the link and entering credentials can result in a compromised mailbox, with the risk of CEO fraud, internal phishing, forwarding rules and access to SharePoint and Teams. Because the links run through legitimate email services, checks that only look at the visible URL miss the real destination.
What should IT administrators do now?
Search the tenants you manage for emails from the three sender domains and with the subject lines and sender names listed, and delete any that were delivered. Check click data and proxy logs to see whether anyone visited reallifeaccounting.com or the .sbs domains. For a suspected victim, reset the password, revoke sessions and refresh tokens, and check MFA registrations, OAuth consents and inbox and forwarding rules. Let users know that outages are only reported in the Microsoft 365 admin centre, never via an email with a restore button.
Copeland refrigeration controllers attacked within two hours of disclosure of 14 flaws, no patch coming
| Vulnerability | 14 vulnerabilities in the Copeland (Dixell) XWEB EVO controller for refrigeration and freezer systems. Nine are in older firmware up to and including 5.2.3.STANDARD and provide, among other things, root access without logging in; five are also present in the latest version 5.8.8.STANDARD, including a predictable daily administrator password |
| CVE number | CVE-2026-124002 to CVE-2026-124015 inclusive; attacked: CVE-2026-124014 and CVE-2026-124010 |
| Actively exploited | Yes, exploitation attempts against CVE-2026-124014 since 12 November 22:00 UTC and against CVE-2026-124010 since 13 November; both added to VulnCheck KEV. According to VulnCheck, the attackers are opportunistic |
| Patch available | No. The XWEB EVO range has been end-of-support since 16 March 2026 and Copeland is no longer releasing fixes. Advice: do not connect to the internet and migrate to XWEB Pro |
How can the vulnerability be exploited?
Researcher Valentin Lobstein of VulnCheck found, among other things, a CGI script that lets you execute commands as root with a single request without logging in, a way to overwrite the system configuration without logging in, a hardcoded administrator password and a daily password that can be calculated from the MAC address and the date. The device itself leaks both the MAC address and its clock time. According to VulnCheck, the first attack request arrived roughly an hour and forty minutes after the CVEs were published, without any public exploit being available. SecurityWeek reports 125 requests from a single IP address on the first night. Security.nl points out that such controllers were also targeted in the recent attack on French supermarkets; VulnCheck sees no link between that campaign and the current attackers.
What could the consequences be?
According to Onyphe figures cited by VulnCheck, almost 2,000 XWEB EVO controllers are exposed on the internet, and most of them are still running the old, most severely affected firmware. An attacker with root access can change temperatures and alarms, resulting in spoiled stock or failure of cold rooms. According to Dutch IT trade publication AG Connect, damage from this kind of refrigeration hack is moreover often not covered by insurance.
What should IT administrators do now?
Find out whether customers in retail, hospitality, food or logistics have XWEB EVO controllers, including those installed by a refrigeration contractor. Take them off the internet and place them behind a VPN in a separate network segment. Search the logs for unauthenticated requests to config.lua and /cgi-bin/xwconsole.cgi and for traffic from 202.61.138.117 and 177.4.12.11. As no further updates will be released, replacement with a supported model is the only lasting solution.
This article was originally published in Dutch on MSP Business. Read the original →
