The United States Cybersecurity and Infrastructure Security Agency has urged Java users to immediately patch a high-severity OpenJDK vulnerability named Java4Shell by Calif.io that attackers can use to crash the Java virtual machine or achieve unauthenticated remote code execution.
Oracle disclosed the flaw, formally CVE-2026-114938, in its November patch dump, which contained 655 patches, 62 of them rated critical, affecting 62 products. Although Oracle did not single out the Java bug for special attention, Calif.io, which discovered the flaw, rated it 9.2 out of 10 under the CVSS 4.0 scheme.
The vulnerability sits in OpenJDK's Base64 decoder, a widely used routine for encoding binary data as text in authentication protocols, file uploads and network messages. Calif.io said the bug lies in optimised machine-level code that uses specialised x86-64 CPU instructions to process many bytes at once. “We discovered a bug in that optimized machine-level code that lets malformed input overwrite neighboring memory. With careful exploitation, the corruption can reveal sensitive memory, as Heartbleed did, or lead to remote code execution. The corruption itself can be triggered by just 64 A characters followed by a single =,” the company said.
Calif.io, which also helped find the PHP2Shell flaw, said it had tested the flaw and found that applications including Keycloak, Atlassian’s Jira and Confluence, Jenkins, Apache Kafka Broker, and Elasticsearch are susceptible to remote code execution. The firm reported that news of the problem had leaked before public disclosure and Oracle’s release of fixes.
Security firm SentinelOne said in an analysis that the flaw “sits in the Java standard library, so exploit traffic is specific to each application.” Its analysis said any Base64 value an application passes to java.util.Base64.Decoder can carry the payload, including HTTP Basic credentials, SAML artifacts, CSRF tokens, and MIME encoded words. In default configurations of the affected applications listed above, attackers can reach the code path prior to authentication, it said.
Cisco’s Talos security team said the flaw is present in OpenJDK Java versions 17.0.6 and later running on x86-64 CPUs with AVX-512 VBMI support enabled. Talos said it had not observed in-the-wild exploitation yet but anticipated probing attempts may pick up soon after public disclosure. It urged adoption of Oracle’s fixed versions but noted it is possible to mitigate the flaw on x86-64 machines by starting the JVM with AVX-512 support disabled.
Akamai’s Chief Information Security Architect Jan Schaumann, writing on the Openwall project’s mailing list, expressed a little discomfort with Calif.io’s work, saying “Calif.io set their priorities on publicity-maxxing, it seems,” pointing out that the firm’s announcement opened by stating that after PHP2Shell they decided to target something more familiar: Java. Calif.io said it burned through just $200 worth of tokens to find Java4Shell.
Analysis
Why This Matters
- The flaw (CVE-2026-114938) affects OpenJDK, a runtime present on millions of servers and endpoints, and can be triggered pre-authentication in widely deployed applications like Jira, Confluence, Jenkins and Elasticsearch.
- The vulnerability’s memory-corruption mechanism resembles Heartbleed, potentially allowing both data leakage and remote code execution with just 64 characters of input.
- The ease and low cost of discovery, $200 in AI tokens, signals that automated tooling is making critical find-and-exploit cycles faster, raising the urgency for proactive patching.
Background
OpenJDK is the open-source reference implementation of the Java Platform, Standard Edition, used in enterprise software, web servers, and development environments worldwide. Base64 encoding is a standard method for representing binary data as ASCII text, commonly used in authentication headers, session tokens, and email attachments. Oracle releases quarterly security patches, and CISA routinely issues binding operational directives for known exploited vulnerabilities in US government systems.
What to Watch
- Whether proof-of-concept or in-the-wild exploitation emerges in the weeks following the public disclosure.
- The adoption rate of fixed Java versions by enterprises running affected applications, particularly those that cannot immediately disable AVX-512 support.
- Further vulnerability discoveries using similar AI-assisted approaches, which may pressure software vendors to accelerate patch cycles.
Checked
12 claims in this story checked against 1 source: 12 supported.
Show each claim
- SupportedCISA urged Java users to immediately patch the high-severity flaw.www.theregister.com - Articles: “The United States Cybersecurity and Infrastructure Security Agency (CISA) has urged Java users to immediately patch a high-severity flaw in the software”
- SupportedCalif.io said it spent $200 in tokens to find Java4Shell, in the context of AI-powered flaw finders.www.theregister.com - Articles: “Such is life in the age of AI-powered flaw finders – which Calif says burned through just $200 worth of tokens to find Java4Shell.”
- SupportedCalif.io named the vulnerability Java4Shell.www.theregister.com - Articles: “Calif had already tested the flaw, which it named “Java4Shell,””
- SupportedAttackers can use the flaw to crash the Java virtual machine or achieve unauthenticated remote code execution.www.theregister.com - Articles: “attackers can use to crash the Java virtual machine or achieve unauthenticated remote code execution (RCE).”
- SupportedOracle's November patch dump contained 655 patches, 62 rated critical, affecting 62 products.www.theregister.com - Articles: “Oracle revealed the flaw in its monthly patch dump for November, which offers 655 patches, 62 of them rated critical, affecting 62 products.”
- SupportedCalif.io rated the flaw 9.2 out of 10 under CVSS 4.0.www.theregister.com - Articles: “Calif applied the CVSS 4.0 bug-rating scheme and gave the flaw a critical-rated score of 9.2/10.”
- SupportedThe story's direct quote about malformed input overwriting memory and potentially leading to remote code execution matches Calif.io's quoted words.www.theregister.com - Articles: ““We discovered a bug in that optimized machine-level code that lets malformed input overwrite neighboring memory. With careful exploitation, the corruption can reveal sensitive memory, as Heartbleed did, or lead to remote code execution.”
- SupportedCalif.io tested the flaw and found Keycloak, Jira, Confluence, Jenkins, Apache Kafka Broker and Elasticsearch susceptible to remote code execution.www.theregister.com - Articles: “Calif had already tested the flaw, which it named “Java4Shell,” and found applications including Keycloak, Atlassian’s Jira and Confluence, Jenkins, Apache Kafka Broker, and Elasticsearch are susceptible to RCE.”
- SupportedCalif.io reported that news of the problem had leaked before public disclosure and Oracle's release of fixes.www.theregister.com - Articles: “Calif.io, which discovered the flaw, used its X account to claim news of the problem had leaked before public disclosure and the release of Oracle's fixes.”
- SupportedSentinelOne said Base64 values passed to java.util.Base64.Decoder can carry the payload, and that affected applications can expose the code path before authentication in default configurations.www.theregister.com - Articles: “Any Base64 value an application passes to java.util.Base64.Decoder can carry the payload, including HTTP Basic credentials, SAML artifacts, CSRF tokens, and MIME encoded words. In default configurations of the affected applications listed above, attackers can reach this code path prior to authentication.”
- SupportedCisco Talos said the flaw affects OpenJDK Java versions 17.0.6 and later on x86-64 CPUs with AVX-512 VBMI support enabled.www.theregister.com - Articles: “the flaw is present in OpenJDK Java versions 17.0.6 and later running on x86-64 CPUs with AVX-512 VBMI (Vector Byte Manipulation Instructions) support enabled.”
- SupportedTalos said it had not observed in-the-wild exploitation and anticipated probing attempts might pick up soon after public disclosure.www.theregister.com - Articles: “has not observed in-the-wild exploitation of this vulnerability yet. However, we anticipate probing attempts may pick up soon after public disclosure.”