Back Mezha Java4Shell flaw could enable remote code execution in widely used Java apps | Межа
Oracle released patches on November 17. Calif recommends temporary JVM options for systems that cannot be updated immediately.
A critical OpenJDK vulnerability dubbed Java4Shell could let unauthenticated attackers remotely execute code on applications including Jenkins, Atlassian Jira, Confluence and Keycloak. Tracked as CVE-2026-114938, the memory-corruption flaw can also expose sensitive information from memory.
This is reported by BleepingComputer .
Calif rated the flaw critical, with a CVSS 4.0 score of 9.2. Oracle gave it a lower severity rating of 8.1 (High) using CVSS 3.1.
Java4Shell affects OpenJDK 17.0.6 and later, along with Java distributions built from the affected code, including Oracle Java SE, Amazon Corretto, Eclipse Temurin, Microsoft Build of OpenJDK, Azul Zulu and Red Hat OpenJDK.
The flaw is in Java’s hardware-accelerated Base64 decoder, which processes encoded data often used in authentication, file uploads and network communications. Calif says specially crafted Base64 data can overwrite portions of Java’s memory. The researchers say the corruption can be triggered with 64 “A” characters followed by a single equals sign (=); older OpenJDK 17 versions require 128 “A” characters followed by the sign.
The vulnerable decoding function is used only on certain x86-64 processors that support AVX-512 VBMI. Calif says the researchers exploited the flaw to achieve remote code execution in several widely used Java applications running with their default configurations, and in sample applications built with Apache Struts and Spring Boot. The researchers said there is no universal exploit: attackers need an application-specific way to deliver malicious Base64 data to the vulnerable decoder.
Calif says researcher Quang Luong found the flaw with help from OpenAI’s Codex AI coding assistant during an audit of OpenJDK’s internals. According to the researchers, Luong found it in one hour and developed an initial working exploit in another two hours, at a total AI usage cost of less than $200. The company says AI agents also helped develop additional application-specific exploitation chains.
Calif published a proof of concept demonstrating the memory corruption but did not release the weaponized exploits used against the affected enterprise applications.
Oracle released patches for CVE-2026-114938 on November 17 as part of its November Critical Security Patch Update, which addresses 655 security vulnerabilities across its product portfolio. The company recommends applying the available patches as soon as possible.
For systems that cannot be updated immediately, Calif recommends disabling the vulnerable hardware-accelerated Base64 functionality by launching OpenJDK with the -XX:-UseBASE64Intrinsics option. Administrators running Java on x86-64 systems can instead use -XX:UseAVX=2 to disable AVX-512 instructions. These are temporary mitigations, and all developers should update Java runtimes to the latest version.
Don't miss other news:
A Pentagon personnel system breach exposed unencrypted personal records of 2.8 million living people and nearly 300,000 deceased, including Social Security numbers and military service details.
The full story
This article is one source in a clustered incident — the cluster page carries the summary, timeline and every other outlet covering it.
