Fake cPGuard Update Backdoored About 2,950 Servers

A security product built to protect web hosting servers was itself turned into a delivery truck for malware. OpsShield, an Indian company that makes cPGuard, a protection tool for Linux web hosting servers, says attackers used stolen credentials to push a malicious update to customers. Its advisory puts the start at 02:06 UTC on November 2, 2026, which was 9:06 PM ET on Sunday, November 1. SecurityWeek reported the company's incident notice on November 4, 2026. We checked the reports on November 5, 2026, at about 7:40 AM ET.

What happened

According to OpsShield's notice, as summarized by SecurityWeek, the intrusion began at the company's billing and licensing server. The attackers allegedly signed up as a customer, ordered a free cPGuard trial to study how the product talked to the licensing system, and then sent crafted requests to the portal until they could run commands on that server.

From there, the company says, they used a stolen access key to upload a fake update to its update distribution network. cPGuard versions 5.88.12 through 5.88.19 contained the malicious code. The poisoned updates were available for roughly 70 minutes, and the attack was discovered within 12 hours, SecurityWeek reports.

Because cPGuard runs with full administrator ("root") rights, any server that checked for updates during that window ran the script with the same power. OpsShield says the script collected data from customer servers, uploaded it under a compromised account and planted a hidden root shell, a secret back door, along with extra files in a temporary folder. The attackers also copied three of the company's cloud servers. Reports differ on where the stolen files went: OpsShield first said they were parked on its own licensing server, but The Hacker News reports the later disclosure says they were sent to a server the attacker controlled.

The timeline, per the vendor

OpsShield's advisory says an "unidentified attacker" used compromised credentials to reach its content delivery network, and that it is still investigating how those credentials were obtained. Help Net Security reports the fake versions were uploaded between 02:04 and 02:09 UTC, and cPGuard agents then installed them as root for more than 70 minutes. Because servers check for updates at intervals, some never ran the script, the advisory notes, and an affected version in a server's update log shows the update was installed but does not by itself prove data left that machine.

The advisory lists the file types the script gathered: Linux account and password-hash files, the root database configuration file, root shell history, and root SSH authorization and known-hosts files. OpsShield cautions that not every file was necessarily taken from every server. The Hacker News, citing the later disclosure, says servers installed the script between 02:08 and 03:20 UTC, and that on a small number of machines it also read cloud storage keys.

How big is it?

OpsShield estimates that about 2,950 servers executed the fake update, according to SecurityWeek. It says the attackers retrieved 32 collected files, each from a different customer server, and copied roughly 7,900 client account records containing names, company names, addresses, email addresses and hashed passwords.

The Canadian Cyber Security Journal's November 4 brief, citing The Hacker News and SecurityWeek, describes the stolen server data as including password hash files, SSH keys and shell history.

What OpsShield says it has done

The company says it removed the attackers from every server they may have touched, replaced SSH keys and passwords across its systems, reset all billing portal sessions, rotated related API keys and is reissuing every cPGuard license key pair because those were compromised too. It is rebuilding its billing, licensing and internal management servers. The advisory first named version 5.88.25 as the fix. The Hacker News reports a later disclosure says 5.88.26 removes the back door and a follow-up update removes a scheduled task the script left behind, so administrators should install the newest release and confirm those files are gone. The company has also published indicators of compromise. OpsShield says it is not recommending a full operating-system rebuild as a routine response, based on what it knows so far.

OpsShield's own guidance, per SecurityWeek: on affected servers, treat all password hashes, root SSH keys, MySQL root passwords and shell history as compromised and change them. "If cPGuard was not online during that window, your server did not receive the fake update. You should still reset your billing portal password," the company says. The Hacker News adds that OpsShield is asking all customers to turn on two-factor sign-in for that portal, and says a server the attacker logged in to by hand should be rebuilt from a clean backup taken before November 1.

Why this matters

This is a supply chain attack: instead of breaking into thousands of servers one by one, the attackers allegedly broke into one trusted vendor and let its normal update system do the spreading. The poisoned files arrived through the same channel customers relied on for genuine fixes.

Our advice (general guidance): if a company runs your website's server for you, send its support team one question: did any of its machines run cPGuard during the November 2 window, and what has it changed since?

The Hacker News reports the attacker took billing records for about 7,900 OpsShield customer accounts, including names, company names, postal and email addresses, and hashed passwords.

Our advice (general guidance): records like those can be reused to make fake vendor messages look personal. If a note about your cPGuard license or account arrives, check it by signing in through a bookmark you saved earlier, never through the message itself.

Our advice (general guidance): server administrators should follow OpsShield's published steps in full, including changing credentials, rather than stopping at the software upgrade.

Questions about a suspicious update or message? Contact us or use our report page.

Sources

Published November 5, 2026 by FraudMeNot Team. Label: Real, reported.

Disclaimer. General education only, not legal, financial or security advice. No guarantee. Consult a qualified professional. News summaries are based on the sources listed and may change as more is reported.

Source: OpsShield incident notice advisory and as reported by SecurityWeek, The Hacker News, Help Net Security and the Canadian Cyber Security Journal, checked Nov 5, 2026, about 7:40 AM ET. FraudMeNot is not affiliated with OpsShield or any outlet cited. General education only, not legal advice, and no guarantee of protection.