Pull to refresh
Logo
Java4Shell flaw enables pre-auth code execution on Java servers

Java4Shell flaw enables pre-auth code execution on Java servers

New Capabilities

A memory-corruption bug in OpenJDK's Base64 decoder affects Jira, Jenkins, Keycloak and other apps on Java 17 and later, when running on chips with AVX-512 VBMI.

Overview

Updated 1 hour ago

A 65-character string can corrupt a Java server's memory before it checks whether you are logged in. The string is 64 capital A's followed by an equals sign, and it breaks the Base64 decoder in OpenJDK, the runtime underneath a large share of corporate Java software.

Calif, the security firm that found the bug, demonstrated remote code execution in Keycloak, Atlassian Jira, Confluence, JFrog Artifactory, Jenkins and PingFederate with default settings. Oracle and OpenJDK released patches on Nov. 17; CISA urged organizations to apply them immediately.

Why it matters

If attackers can reach a login form on an unpatched Jira or Keycloak server, they can run code without an account.

Questions about this story

Free account needed to ask — your question is kept and asked for you right after sign-up. Answers are public.

No questions yet — be the first to ask.

Key Indicators

9.2
CVSS 4.0 severity score
Calif rates the flaw 9.2 out of 10, in the critical range.
6
Apps with pre-auth RCE demonstrated
Keycloak, Jira, Confluence, Artifactory, Jenkins and PingFederate, in default configurations.
65
Characters in trigger string
64 'A' characters followed by a single '='; older OpenJDK 17 builds need 128.
62
Bytes written past buffer
The decoder writes 62 bytes beyond a 50-byte output array.
90%
Fortune 500 companies running Java
Calif's estimate of Java's reach in large enterprises.
0
Confirmed in-the-wild exploits
As of Nov. 17, no attacks reported; the CVE is not in CISA's KEV catalog.

Voices

Curated perspectives — historical figures and your fellow readers.

Sign up to generate historical perspectives on this story.

People Involved

Organizations Involved

Timeline

4 events Latest: Yesterday
Tap a bar to jump to that date
  1. Calif discloses Java4Shell

    Latest Disclosure

    Calif publishes CVE-2026-114938 and demonstrates pre-auth RCE in six widely used Java applications.

  2. Oracle and OpenJDK release patches

    Patch

    Fixes ship for Java 17, 21, 25 and 27; Amazon updates Corretto while Red Hat prepares its own.

  3. CISA urges immediate patching

    Government

    CISA tells organizations to inventory Java installations and prioritize internet-facing systems.

  4. Independent labs reproduce crash and leak

    Research

    Upwind reproduces corruption on an Ice Lake server; Blackpoint crashes a Spring Boot app and reads a nearby secret.

Scenarios

1

Public exploit code lands and CISA adds Java4Shell to its exploited-vulnerabilities catalog

Possible Resolves by Jan 31, 2027

Discussed by: Security researchers tracking the CVE, including Rapid7 and Qualys.

A working exploit for one of the pre-auth targets circulates publicly, or attackers begin exploiting it in the wild. CISA's Known Exploited Vulnerabilities catalog is the trigger: once a CVE is listed, US federal agencies face a patching deadline and the flaw becomes a priority for everyone else.

2

Vendors patch fast enough that mass exploitation never materializes

Likely Resolves by Feb 28, 2027

Discussed by: CISA and Calif, both of which emphasized patching and compensating controls.

Oracle, OpenJDK, Amazon and Red Hat ship fixes within weeks. Internet-facing Jira, Jenkins and Keycloak servers get patched or shielded by proxies that buffer request bodies. The flaw remains a risk for bundled JDKs and container images, but no widespread campaign appears.

3

Attackers use JavaBleed to steal secrets without achieving code execution

Possible Resolves by Q1 2027

Discussed by: Blackpoint Cyber's Adversary Pursuit Group, which reproduced the heap read in 30 of 30 local runs.

Instead of chasing RCE, attackers trigger the overflow to corrupt an array length and read back JVM heap memory. Private keys, tokens and personal data sit in that memory. A single leaked credential can be enough to move laterally, even if the full exploit chain never works.

Historical Context

2 moments from history that rhyme with this story — and how they unfolded.

April 2014

Heartbleed (2014)

A missing bounds check in OpenSSL's heartbeat extension let attackers read up to 64 kilobytes of server memory per request. The bug affected a large share of HTTPS servers and exposed private keys, session cookies and passwords.

Then

OpenSSL released a fix within days; companies rushed to patch and revoke certificates.

Now

Heartbleed became a reference point for memory-safety bugs that leak secrets rather than execute code.

Why this matters now

JavaBleed is the same class of bug: an out-of-bounds read that exposes heap memory before any code execution.

December 2021

Log4Shell (2021)

A lookup feature in Apache Log4j 2 let attackers run code by sending a crafted string to a logged field. The library was embedded in thousands of Java applications, and exploitation began within hours of disclosure.

Then

Organizations patched in waves; attackers scanned for vulnerable servers and deployed cryptominers and backdoors.

Now

The incident pushed software bills of materials and dependency scanning into mainstream security practice.

Why this matters now

Java4Shell follows the same pattern: a flaw in a low-level Java component that surfaces in many products at once, with pre-auth reach through login forms.

Sources

(11)