Artificial Intelligence

Java4Shell Turns Base64 Into Pre-Auth RCE

Calif says Java4Shell turns a Base64 decoding bug into unauthenticated remote code execution in major Java apps.

StartupHub.ai Staff

By StartupHub.ai Staff, reviewed by Daniel Singer, Editor

2 min read

Java server code under review for Base64 memory bug
Credit: Calif

Calif says Java4Shell achieves unauthenticated remote code execution against Keycloak, Jira and Jenkins in default configs.

The issue is tracked as CVE-2026-114938 and scores 9.2 under CVSS 4.0, according to Calif.

Aikido reports the bug affects Java 17, 21, 25 and 27 releases before patched builds.

Base64 is a text encoding that carries binary data through login tokens, uploads and network messages.

Java promises memory safety where the runtime manages memory and blocks classic corruption bugs.

It breaks that promise here.

An intrinsic is a JIT shortcut that replaces Java code with hand tuned CPU vector instructions for speed.

Calif says the vulnerable AVX-512 path mishandles the final characters and computes a negative output length.

The trigger is 64 A characters plus a single equals sign, which allocates a 50 byte buffer then writes 64 bytes, according to Calif.

That stray padding turns the write mask to all ones so 62 bytes spill past the array into neighboring objects, Calif reports.

Quang Luong found the bug in an hour and built the first exploit in two more for under 200 dollars in tokens, according to Calif.

Ordinary login and file upload paths expose the decoder to remote input, according to Calif.

Researcher samczsun showed a malicious TLS server can also compromise a Java client that connects out, according to Calif.

Calif also reports post-auth execution in Kafka, Solr and Elasticsearch plus Struts and Spring Boot samples.

The fixed overwrite becomes a writer that corrupts a buffer length, then a reader that leaks heap memory like Heartbleed, according to Calif.

A TLAB is a private Eden slice that gives each thread its own cursor, and grooming forces large arrays outside it to order victims, according to Calif.

There is no universal exploit because each app needs different grooming and repair steps, according to Calif.

Patch to fixed builds now and you can also watch for needless native optimizations that widen attack surface.

Sources

Get the free daily brief every morning

Every morning the day's AI news, new AI startups and new funding rounds. Free and takes a few minutes to read, unsubscribe at any time.

Join 64,000 other subscribers.

TopicsCalifOpenJDKKeycloakJiraJenkinsQuang LuongsamczsunAVX-512

StartupHub.ai StaffDaniel SingerWritten by StartupHub.ai Staff, Editorial teamReviewed by Daniel Singer, EditorAbout

StartupHub.ai Staff, Editorial team

The staff writers of StartupHub.ai, ranging from investment analysts to avid AI tool users, early adopters and critical enthusiasts. Backgrounds span engineering, business and the arts. We hold every piece to rigorous standards of research and review.

Daniel Singer, Editor

Daniel Singer is an AI expert who has been writing about, covering and building AI technology since 2018.

More from Daniel Singer

Test your Agent Readiness now

Paste your website. How well ChatGPT, Claude and the other agents can read and use a site, in about a minute.

Watch

The videos behind the news, one flip at a time

The feed mixes the videos from our articles with startup demos and you flip through them like TV channels for free with no signup.