The cyber attacks this week, 2 to 8 November 2026, were defined by attackers coming in through someone the victim trusted rather than through the victim’s own front door. An attacker who broke into OPSSHIELD’s billing server pushed a malicious cPGuard update that about 2,950 customer servers installed on 2 November, and Crédit Mutuel Arkéa said attackers entered through a service provider’s platform and took about €100 million. France’s blood service EFS reported tampering with plasma freezers through a refrigeration supplier, the latest step in refrigeration attacks that began in French supermarkets in October, and Google disclosed on 2 November that it had disrupted a botnet built into the firmware of more than 1.3 million TV boxes and other Android devices. Myriad Genetics told the SEC it suspended testing operations after a network intrusion, and DreamHost’s breach filing showed that its File Manager had exposed customer FTP and SFTP logins since August. In five of the six incidents the entry point was a vendor, provider, supplier, manufacturer or web host, and Myriad has not disclosed how its intruder got in.
1. OPSSHIELD cPGuard update channel hijacked to backdoor about 2,950 hosting servers
Overview: On 3 November 2026, OPSSHIELD, an Indian security software company whose cPGuard agent protects Linux web hosting servers, disclosed that an attacker who had broken into its billing server used a stolen access key to publish a malicious cPGuard update, which about 2,950 customer servers installed on 2 November. This incident disclosure is the later and fuller of OPSSHIELD’s two notices; its earlier update-channel advisory points readers to it for the latest recommended actions.
Impact: OPSSHIELD says the fake update, delivered as cPGuard versions 5.88.12 to 5.88.19, shipped for about 70 minutes on 2 November, and that about 7,900 billing portal accounts were copied, carrying names, company, postal address, email and bcrypt-hashed passwords. The attacker also made copies of 3 OPSSHIELD cloud servers and retrieved an estimated 32 of the files the update collected, each tied to a customer server. An OPSSHIELD spokesperson gave BleepingComputer a figure of 2,946 servers. This is a supply chain incident rather than a software flaw, so there is no CVE, CVSS score or KEV listing to track it by.
Details: OPSSHIELD says the attacker registered a customer account under a false identity, ordered a free trial and, late on 1 November, sent crafted requests to the billing portal’s callback pages that gave it command execution on the billing and licensing server; its logs do not record the exact request. With a stolen access key, the attacker uploaded the fake update between 02:04 and 02:09 UTC on 2 November, and agents installed it between 02:08 and 03:20 UTC. The update ran as root and copied /etc/shadow, which holds the password hashes of every local account, along with root SSH keys, the MySQL root password where present, and root shell history, which often contains typed passwords and tokens. The billing-server foothold is consistent with MITRE ATT&CK T1190 (Exploit Public-Facing Application), followed by T1078 (Valid Accounts) for the stolen key and T1195.002 (Compromise Software Supply Chain) for the poisoned update; the attacker has not been identified.
Remediation guidance:
/etc/cron.d/i93677, /var/tmp/.i93677/ and /var/tmp/.cpg91511_sh.2. Crédit Mutuel Arkéa loses about €100 million through a service provider’s platform
Overview: On 4 November 2026, French banking group Crédit Mutuel Arkéa announced it had been the victim of a cyberattack in which attackers entered through a service provider’s platform and carried out transactions from the bank’s own account, and on 5 November the bank said about €100 million had been stolen, Reuters reported. Arkéa holds about €211.5 billion in assets and serves more than 5.8 million members and customers.
Impact: On 2 November, blockchain investigator ZachXBT flagged more than $106 million in suspicious withdrawals, later raising his estimate above $133 million and naming the bank, according to CoinDesk and The Block. The funds were withdrawn first as USDC and USDT stablecoins and then swapped onward, including into Monero. On 4 November the bank declined to confirm a reported €118 million figure, saying it could only validate the order of magnitude, and on 5 November it put the loss at about €100 million. Arkéa says part of the funds has been blocked, and on 6 November Daniel Baal, president of the Confédération Nationale du Crédit Mutuel, said on France 5 that Arkéa had already recovered part of the money, according to Boursorama. The bank says customer accounts were not used and that the attackers did not enter through its online bank Fortuneo. It told Reuters it had informed the European Central Bank, French banking supervisors and prosecutors, and the Paris prosecutor’s cybercrime unit has opened an investigation.
Details: Arkéa says the attackers went through “the platform of one of our service providers” and then reached the bank directly, using the bank’s own account to carry out fraudulent instant transfers to other banks, which it says may then have let them convert the money into cryptocurrency. Instant payments settle in seconds and are difficult to recall, which is part of why recovery is partial. Reporting indicates the target was the bank’s treasury rather than customer balances. The timeline carries an uncomfortable detail: on 27 October, a week before the theft, Arkéa detected signs that a service provider’s systems had been compromised, cut the connection and notified the authorities, but at that point nobody connected that compromise to Arkéa itself. The provider has not been named. Asked whether human error caused the attack, the bank answered: “No. This was a sophisticated attack from outside.” Arkéa chairman Julien Carmona has said the attackers were very likely assisted by advanced AI, though the precise role of AI remains unconfirmed. The entry route maps to MITRE ATT&CK T1199 (Trusted Relationship). No threat actor has been attributed, and ZachXBT said he had a high degree of confidence the attacker is not linked to North Korea’s Lazarus Group.
Remediation guidance:
3. Refrigeration attacks escalate from supermarkets to France’s blood service
Overview: The refrigeration attacks escalated in three steps. On the night of 11 to 12 October 2026, refrigeration in hundreds of French supermarkets heated up overnight. On Monday 2 November, fridges and freezers at more than 100 shops across Belgium, Luxembourg, Sweden, Ireland, Estonia and the Netherlands ran hot. On 3 November, Établissement Français du Sang (EFS), France’s blood service, confirmed that attackers had tampered with plasma freezers at three of its sites through a refrigeration supplier, in an attack Le Canard enchaîné dated to the night of 1 to 2 November.
Impact: Le Canard enchaîné, as reported by IT-Connect, said freezers normally kept around minus 30°C had been reprogrammed to a setpoint of 110°C; that is the target the attackers set, not a temperature the plasma reached. An on-call technician spotted the abnormal temperature rise during the night, and staff moved blood products to units not connected to the compromised platform. EFS initially described the affected material as donor samples kept for transfusion monitoring, then told Le Canard enchaîné that frozen plasma intended for patient transfusion was also involved. EFS says no health consequences have been identified, no blood products were lost, and blood collection and distribution continue, according to the Associated Press. In the European wave, Carrefour said about 15 of its 700 Belgian stores were hit. On 5 November 2026, CISA, the FBI, HHS, USDA and DOT published a joint fact sheet urging operators to harden industrial refrigeration and cold-chain monitoring systems.
Details: EFS’s own statement referred to a supplier of refrigeration equipment without naming it; press reporting identifies the supplier as Coolworld, a Dutch refrigeration services company, which confirmed a cybersecurity incident affecting a limited number of its units and involving third-party remote-monitoring software. Investigators have not publicly identified the attackers or how they gained access, and France’s national cybersecurity agency and Health Ministry are investigating. The Paris prosecutor has opened a formal inquiry into the supermarket attacks, describing a large-scale cyberattack on connected refrigeration equipment. The US fact sheet says attackers may exploit internet-accessible or unpatched temperature and refrigeration controllers through authentication bypass, arbitrary code injection and weak configurations, then force continuous heating and defrost cycles. The EFS activity is consistent with that description and with MITRE ATT&CK for ICS T0883 (Internet Accessible Device) and T0836 (Modify Parameter). One caveat on the escalation narrative: French reporting notes that nothing yet establishes the equipment Coolworld manages as being among the controller models named in the CERT-FR alert on the supermarket wave, or the attackers as the same people. Awais Rashid, professor of cybersecurity at the University of Bristol, told Al Jazeera on 3 November: “There are known vulnerabilities in several controllers (many of which are past their end of life).” ANSSI has published guidance against these configurations since 2014.
Remediation guidance:
4. Myriad Genetics 8-K: network intrusion suspends genetic testing operations
Overview: On 5 November 2026, Myriad Genetics filed a Form 8-K under Item 8.01 disclosing that on 2 November it detected unauthorized access within its information technology network and temporarily suspended its testing operations and order management systems.
Impact: Myriad says most systems have since returned to normal operation. It is assessing whether any patient records were accessed on systems that contain patient information, and it has not published any count. The company says it does not believe the incident will have a material adverse effect.
Details: Myriad has not disclosed the entry point, the malware involved or any threat actor, so no ATT&CK mapping is supported. The filing is under Item 8.01 rather than Item 1.05, consistent with the company’s statement on materiality. Whether patient data was taken remains open, and it is the question that will decide notification duties.
Remediation guidance:
5. Google disrupts XCPE botnet built into the firmware of 1.3 million TV boxes
Overview: On 2 November 2026, Google Threat Intelligence Group reported that it had disrupted the XCPE botnet, a backdoor preinstalled in the firmware of Android TV set-top boxes made by Videostrong Technology Co.
Impact: GTIG says more than 1.3 million set-top boxes were enrolled in the botnet, and assesses that taking down the affiliates’ command domains will sever access to at least 10 million additional compromised devices. Most infected devices were Android TV boxes, but GTIG also observed infected telehealth devices, karaoke machines and digital signage devices, which takes the botnet beyond living rooms into clinics, venues and offices. Access was sold for residential proxies and ad fraud to more than a dozen malware families and affiliates. Google took legal action against the command servers and affiliate domains, and Videostrong is banned from building certified Android devices.
Details: GTIG says the backdoor sits in the firmware partition, runs with system privileges, reinstalls itself if removed, disables Google Play Protect and takes commands over MQTT. GTIG assesses with high confidence that the backdoor was developed and operated for financial gain by Videostrong itself, and says Videostrong used obfuscation to load malicious code after first boot, exploiting a gap in Google’s tests. Partners in the disruption included Japan’s National Police Agency, Lumen, HUMAN Security, Plume, Synthient and TrendAI. The preinstalled backdoor maps to MITRE ATT&CK for Mobile T1474 (Supply Chain Compromise).
Remediation guidance:
6. DreamHost File Manager breach exposed FTP and SFTP logins from August to October
Overview: DreamHost said in an incident notice that on 23 October 2026 it discovered an unauthorized party had accessed the server running its web-based File Manager by exploiting a vulnerability in the third-party software behind it. On 4 November, TechCrunch reported, citing a breach notice DreamHost filed with California’s attorney general, that the attack began in August and that hackers stole customer emails and credentials.
Impact: The attacker could capture the FTP and SFTP usernames and passwords of customers who logged into File Manager between about August 2026 and 23 October, and the login credentials of a limited number of webmail users. In a small number of cases DreamHost believes those credentials were used to view mailbox contents. A DreamHost spokesperson told TechCrunch that well under 10 percent of customers were affected, and DreamHost says it found no evidence that control panel logins or billing details were accessed.
Details: DreamHost has not named the vulnerable software or the attacker. It reset affected passwords and ended active webmail sessions. FTP and SFTP credentials give write access to customer websites, which is why DreamHost tells affected users to check site files for changes they do not recognise and to review webmail forwarding rules and filters. The access is consistent with MITRE ATT&CK T1190 (Exploit Public-Facing Application) and T1056.003 (Web Portal Capture).
Remediation guidance:
Related reading
- Weekly Cybersecurity Intelligence Report: Cyber Threats and Breaches, 26 Oct to 01 Nov 2026: Binance, Amadeus and UNM Health, last week’s single footholds with outsized reach.
- Weekly Cybersecurity Intelligence Report: Cyber Threats and Breaches, 12 to 18 Oct 2026: the French supermarket refrigeration controller attacks that preceded this week’s EFS incident.
- Weekly Cybersecurity Intelligence Report: Cyber Threats and Breaches, 19 to 25 Oct 2026: LiteLLM, Railway and Navigate360, where an intermediary’s credentials opened many downstream victims.
How FireCompass helps
The common entry point this week was a party the victim trusted: a security vendor’s update channel, a bank’s service provider, a refrigeration supplier’s remote-monitoring software, a device manufacturer’s firmware and a web host’s file manager. Two of these were reachable from outside and two were not, and the difference is worth being honest about. A poisoned update signed into your own agent, or firmware shipped from the factory, is an identity and procurement problem rather than an attack surface one. What external testing does answer is the rest: which controllers, portals and vendor-managed systems carry your name and respond on the internet, which of them an attacker can reach something useful through, and whether credentials tied to them are already circulating. FireCompass discovers those assets continuously and validates what is exploitable with proof of exploit, at a false positive rate under 2 percent against 40 to 70 percent from leading DAST tools.
Find the third-party paths an attacker would try first.
Frequently asked questions about cyber attacks this week
What was the biggest cyberattack this week (2 to 8 Nov 2026)?
The Crédit Mutuel Arkéa theft was the largest disclosed financial loss of 2 to 8 November 2026. The French bank said on 4 November that attackers entered through a service provider’s platform and used its own account for fraudulent instant transfers, and on 5 November it said about €100 million was stolen, Reuters reported. Customer accounts were not used, and part of the money has since been recovered.
What happened in the OPSSHIELD cPGuard supply chain attack?
An attacker who broke into OPSSHIELD’s billing server used a stolen access key to publish a malicious cPGuard update that about 2,950 customer servers installed on 2 November 2026. The update ran as root and copied local password hashes, SSH keys, database credentials and shell history. OPSSHIELD names 5.88.26 as the clean release, and affected servers need every credential rotated as well as the leftover files removed.
Was the French blood bank EFS hacked?
Yes, indirectly. Établissement Français du Sang confirmed on 3 November 2026 that attackers tampered with plasma freezers at three sites through a refrigeration supplier on the night of 1 to 2 November. Reporting identifies the supplier as Coolworld, a Dutch refrigeration company, which confirmed an incident involving third-party remote-monitoring software. An on-call technician caught the change, and EFS says no blood products were lost.
Did Crédit Mutuel Arkéa know about the provider compromise beforehand?
Partly. On 27 October 2026, a week before the theft, Arkéa detected signs that a service provider’s systems had been compromised, cut the connection and notified the authorities. At that point nobody publicly connected that compromise to Arkéa itself. The bank’s chairman, Julien Carmona, has said the attackers were very likely assisted by advanced AI, though that role is unconfirmed.
What did Myriad Genetics disclose about its cyberattack?
Myriad Genetics filed an 8-K on 5 November 2026 saying it detected unauthorized access to its IT network on 2 November and temporarily suspended testing operations and order management systems. Most systems have returned to normal, and Myriad is assessing whether patient records were accessed. It does not expect a material adverse effect.
What pattern connects this week’s cyberattacks?
A trusted third party connects five of the six major incidents of 2 to 8 November 2026: OPSSHIELD’s update channel, a service provider’s platform at Crédit Mutuel Arkéa, a refrigeration supplier at EFS, the manufacturer’s firmware in XCPE devices, and DreamHost’s File Manager. Myriad Genetics has not disclosed its entry point.
How can organisations defend against attacks through trusted vendors?
Rank suppliers by the access they hold rather than by contract value, and test those access paths rather than surveying them. In the week of 2 to 8 November 2026, a stolen OPSSHIELD access key let a poisoned update reach about 2,950 servers in roughly 70 minutes, and a service provider’s platform opened a path to about €100 million at Crédit Mutuel Arkéa a week after that provider’s compromise had already been detected.
