A vulnerability identified in a component used by numerous programs has raised alarm among cybersecurity experts. At the center is Java, a technology that has contributed to the functioning of online services and business applications for over thirty years. It often works behind the scenes, on servers, without users realizing its presence.
The flaw, dubbed Java4Shell, was made public on November 17, 2026. The relevant authorities reported the risks and Oracle released the updates necessary to correct it. However, not all systems that use this technology are exposed: for an attack to be possible, very specific conditions must be met.
What is Java and why is it so widespread in computer systems
Java is a programming language born in the 1990s and today used to create business applications, web platforms and tools intended to run on different operating systems. It should not be confused with JavaScript, which, despite the similar name, is another language.
One of its main features is portability: a Java program can run on different computers, as long as they have a compatible environment. What makes this possible is above all the Java Virtual Machine (JVM), a sort of intermediary that allows the device to execute the application’s instructions. Hence the motto “Write once, run anywhere”.
The data from Stack Overflow’s Developer Survey 2026 also gives an idea of its diffusion. According to what emerged, 27.6% of participants who answered the question on programming languages declared having used Java in the last year.
Java is used, for example, in banking software, authentication services and development tools. Many applications work on servers, which are computers that provide functionality to other devices over a network. Some programs include their own version of the Java environment, which is not always easy to locate.
How the Java4Shell vulnerability works and what security risks it poses
The flaw, identified with the code CVE-2026-114938, was discovered by researchers from the company Calif and concerns OpenJDK, an open source version of the Java platform. More precisely, it concerns a function that is used to read information encoded in Base64, a format used to represent data through textual characters, also during the procedures for accessing online services.
The problem occurs when this function receives a specially constructed character sequence. On some processors, in fact, the program can make an error and write data beyond the memory space that was assigned to it. Experts talk about memory corruption: it’s a bit like filling a container beyond its capacity, ending up invading the one next to it and altering its contents. In software, this can cause a crash or, in worse cases, pave the way for a cyber attack.
A cyber criminal could exploit this flaw to read confidential information in memory or, in certain circumstances, execute code remotely, i.e. give instructions to a system even without having logged in. The researchers demonstrated this possibility during tests on applications such as Jenkins, Jira and Keycloak, used by companies to develop software, manage projects and control user authentication.
To exploit the flaw, however, several conditions must occur simultaneously: a vulnerable version of the software is needed, a 64-bit Intel or AMD processor compatible with particular instructions called AVX-512 VBMI and an application that allows data prepared by the attacker to reach the affected function. Systems with ARM64 processors, however, are not affected.
As for OpenJDK versions, the problem affects line 17 from 17.0.6 up to versions prior to 17.0.21.1, as well as some versions of lines 21, 25 and 27, prior to the respective corrective updates. A specific version of Java 8 Enterprise Performance Pack and some GraalVM distributions are also involved. Standard Java 8 and Java 11 do not have this flaw. To understand if a system is exposed it is therefore necessary to check the full version installed.
The severity of the flaw was rated with a score of 8.1 out of 10 according to the CVSS 3.1 scale, used to classify cyber vulnerabilities, as reported by CERT-EU. Calif has also published a Proof of Concept, i.e. a demonstration code that allows you to reproduce the problem, but is not enough on its own to take control of a system. As reported by Tenable, other researchers have instead declared that they have developed tools capable of exploiting it in specific applications.
According to data published by GreyNoise, a cybersecurity company that monitors suspicious activities on the Internet, as of November 19, over 180 requests compatible with attempts to exploit the flaw had already been detected, coming from five IP addresses, without however finding successful intrusions on their sensors.
How to protect systems and prevent cyber attacks
The first measure against Java4Shell is to install the security updates published by Oracle and the OpenJDK project on November 17. For the major affected versions, Java 17.0.21.1, 21.0.13.1, 25.0.5.1, and 27.0.1.1 are listed as correct. For Java 8 Enterprise Performance Pack the 8u513-perf is available. Other suppliers may follow different calendars and numbering, so it is important to consult their respective notices.
To apply fixes, a company must first identify where Java is used by checking servers, cloud services, and applications that include their own version of the software. Once vulnerable systems are identified, priorities can be established. The recommendation is to intervene first on the services accessible from the Internet, which are more easily reachable by attackers, and to restart the updated Java programs for the changes to take effect.
Before deploying a large-scale update, especially on major systems, it is also a good idea to check its compatibility with existing applications and, after installation, check that the fix was applied correctly.
If immediate action cannot be taken, administrators can temporarily disable a Java feature that speeds up processing of Base64 data that is implicated in the vulnerability. This measure may affect performance and is not a substitute for installing the permanent update.
Meanwhile, the speed at which vulnerabilities are discovered is also changing. Back in September, Oracle announced plans to release more frequent updates in 2027 for Java, noting that AI is speeding up problem finding and fixing. The same Calif researchers declared that they used it to study Java4Shell, after having also used similar tools for the discovery of two flaws in the PHP language.
Prevention, however, does not end with this single flaw. Keeping an updated list of programs, eliminating unused ones and periodically checking the installed versions allows you to more easily identify exposed components. Limiting account privileges and separating sensitive systems from the rest of the network can limit damage in the event of an intrusion. Monitoring of suspicious activities and backups, i.e. security copies of data, are also important, along with already verified procedures for restoring services after an incident.
Multi-factor authentication, which requires additional verification beyond the password, is also helpful against credential theft. However, it does not directly prevent the exploitation of Java4Shell, because in some applications the attack can occur before the identity of the person making a request is even checked.
Anyone who uses a site or online platform cannot directly correct a flaw present on the servers: this responsibility lies with the managers. On your computer or smartphone, however, you can reduce risks by keeping the operating system and applications updated.
Calif Stack Overflow CERT-EU CSIRT Tuscany Red Hat Tenable Oracle – Critical Security Patch Update, November 2026 Oracle – Transitioning Java to more frequent security updates GreyNoise – Java4Shell Exploitation Activity Dashboard









