Latest

Cybersecurity

Barracuda ESG security falls with two chained flaws

Two chained vulnerabilities allow an unauthenticated attacker to reach root on Barracuda ESG, and active exploitation is already underway with a public PoC.

Translated from Spanish by an automated system. The Spanish original prevails. Read the Spanish original

Barracuda ESG security falls with two chained flaws
Barracuda ESG security falls with two chained flaws

From no credentials to root on your email filter: the November 16, 2026 alert describes how an attacker chains CVE-2026-116606 (9.8) and CVE-2026-111560 (7.2) in Barracuda ESG to go from the administration interface to root without stealing a password. Both affect Barracuda Networks products, both have a public exploit and both are being used.

The alert on CVE-2026-116606 in Barracuda ESG details the severity of each piece. CVE-2026-116606 scores 9.8 out of 10 and allows an unauthenticated attacker to bypass authorization controls. CVE-2026-111560 scores 7.2 and allows system commands to be executed with root privileges, although it requires prior administrative access.

The interest lies in the combination. The first opens the administration interface door without credentials; the second turns that access into full control of the appliance. An attacker who chains both goes from being outside to being root without having stolen a password, and that is the kind of chain that turns an annoying vulnerability into a full-blown security incident. The question is no longer whether there is a patch, but how many panels remain exposed while it is installed.

Which products are affected?

The list of affected appliances is wider than the product name suggests. It includes Email Security Gateway, ADC / Load Balancer, Message Archiver, WAF, WAF Control Center and Web Security Gateway. They all share the same administration interface and, therefore, the same entry point.

The official recommendation is to upgrade to the latest available version. When that is not immediately possible, the advice is restrictive and sensible: limit access to the administration interface to trusted internal networks and use network access controls to reduce exposure. In practice, that means no longer having the administration panel visible from the internet, which is exactly what almost nobody does when administering from home.

The restriction is not cosmetic. An exposed panel receives authentication attempts from all over the world every day, and the difference between an attempt and an intrusion is a chain like this one.

Neither measure is new or exotic. Both have appeared in every secure configuration guide for a decade, and both are skipped just as often for operational convenience. The underlying problem is that an email gateway is usually administered from anywhere, because whoever maintains it is not always in the same office as the appliance.

The role of an email gateway

An email filter is not a peripheral service. It is the door through which most of the content a company receives from outside enters, and also the point where integration credentials with the rest of the ecosystem are stored. If someone obtains root there, they inherit the appliance's privileged position: they read the email passing through, modify rules, add trusted senders and place themselves at the center of the chain.

That effect is the same one seen with the chain of flaws that compromised thousands of corporate appliances in early November: the value of the attacked appliance is not in its data, it is in its position. Devices that act as intermediaries between employees and external services are the most profitable targets because they are never reviewed the way a production server is.

The calendar does not help either. The alert arrives two weeks after November Patch Tuesday, with 1,094 vulnerabilities fixed by Microsoft, and security teams are working through that backlog. When every vendor publishes at once, what gets delayed is not the difficult patch, it is the one for the secondary application nobody remembers having.

The chain as a 2026 pattern

This year it has stopped being surprising that exploitation chains are the norm rather than the exception. One flaw provides initial access, another performs the escalation, and the craftsmanship goes into joining the pieces. The result is that two medium-severity vulnerabilities can produce a critical incident, and individual ratings do not reflect that.

The case of the OpenSSH bug that put Linux and musl at risk showed the most uncomfortable variant: a flaw announced as limited that, in combination with other conditions, changed category. Classifying a flaw in isolation helps prioritize patches, not predict the real risk of a specific infrastructure.

For the security manager, the consequence is uncomfortable. Severity metrics help decide what gets patched first, but not what actually gets reviewed. A team that only looks at the 10 score overlooks the 7.2 behind it, and that is exactly the one that closes the chain.

What to do? Isolate the panel, patch and look for persistence

The first thing is to know whether the product is on the list and, above all, whether its administration panel is accessible from the internet. An inventory of exposed interfaces resolves more incidents than any state-of-the-art detection tool, and it takes an afternoon with a spreadsheet and a scan.

The second is to update with judgment. Emergency patches sometimes break functionality, and the temptation to wait until the following weekend is human. In this case there are two aggravating factors: a public exploit and ongoing exploitation, which reduces the room for negotiation to hours or days, not weeks.

The third is to review the evidence. If an attacker obtained root on an email gateway, they left traces in transport rules, service accounts and authentication logs. Looking only for the malicious file is enough to close the ticket, not to know whether the access is still live.

The review must include created accounts, new rules and changes to trust lists. An attacker who reaches root does not need to come back in if they leave a door open with their name on it.

What this incident says about the model

A security vendor whose products are used to protect others is, by definition, an appetizing target. Attackers know it and exploit it: every flaw gains value, and every published exploit degrades trust in the entire brand. It is no coincidence that this month's alerts pile up around tools that occupy privileged positions in the corporate network.

What changes over time is speed. Active exploitation and the public proof of concept arrive almost together, and the time between publication and the first attack is now measured in hours. A management model that assumes days of margin is outdated by definition.

The useful takeaway for the reader is just one. Appliances bought for the security label and installed in a corner of the network are the ones that end up on incident lists, because nobody looks at them until the vendor warns. If you expose the administration panel to the internet: restrict it today, patch within 48 hours and review accounts, rules and authentication logs.

José Maldonado

Author

José Maldonado

Activist and blogger on technology, free software and blockchain.

The latest
The price of DDR5 multiplies by 5.5 and hardware is not getting cheaper

Keep reading