Researchers at Calif have disclosed Java4Shell, a memory-corruption flaw in OpenJDK's Base64 decoder that can lead to remote code execution. The bug, tracked as CVE-2026-114938, scores 9.2 out of 10 under CVSS 4.0.

The vulnerability affects the Base64 decoder in OpenJDK, the most popular Java runtime. It uses specialized x86-64 CPU instructions to decode faster, and malformed input can overwrite neighboring heap memory. According to Calif, a remote attacker with no credentials can use this to read sensitive memory, similar to Heartbleed, or to achieve remote code execution. The corruption can be triggered by 64 A characters followed by a single equals sign. On OpenJDK 17, the trigger input is 128 A characters followed by an equals sign.
Calif said its initial testing demonstrated remote code execution in several widely deployed Java applications using their default configurations. Pre-authentication examples include Keycloak, Atlassian Jira and Confluence, JFrog Artifactory, Jenkins, and PingFederate. Apache Kafka Broker, Apache Solr, Elasticsearch, Gerrit, and Spinnaker were affected after authentication. Calif also said that samczsun shared a client-side exploit, in which a Java program connecting to a malicious server over TLS can be compromised even if it runs no service of its own.
Calif used AI tools to find and weaponize the flaw. The company said researcher Quang Luong discovered the vulnerability in about an hour and built a first working exploit in two more hours. Calif also argued that many Java features add code that most applications never need, creating risk for everyone. It proposed a practice it calls vibe-uncoding, which uses AI to identify and remove unneeded code to reduce attack surface.
Source: calif.io







