Three-day deadlines don't leave room
for a missed email.
CyberComply monitors the CISA KEV catalog 24/7 and alerts you within minutes of every new listing, so the clock never starts without you.
KEV Watch: Java4Shell (CVE-2026-114938)
Audience: Federal Contractors · DevOps & Platform Teams · Security Operations Leaders Classification: TLP:CLEAR Reporting Period: November 18, 2026
This is a KEV Watch edition: CVE-2026-114938 is not in the KEV catalog, and no exploitation in the wild has been reported. We are covering it now for the same reason we covered PHP2Shell early. CISA issued an alert on November 17 urging immediate patching, exploit code is already circulating, and the researchers who found it, Calif, are the same team behind PHP2Shell, which went from disclosure to a KEV listing in five days. The time to inventory is before a deadline clock starts.
What's Confirmed, and What Isn't
CVE-2026-114938, nicknamed "Java4Shell," is a heap out-of-bounds write in the HotSpot Java Virtual Machine's Base64 decoder. When an application decodes the same kind of Base64 data often enough, HotSpot's just-in-time compiler swaps the ordinary Java decoder for a hand-written, hardware-accelerated routine. That routine miscounts padding: specially malformed Base64 makes it write 62 or 63 bytes past the end of its output buffer. Oracle and the OpenJDK Vulnerability Group shipped fixes on November 17.
The confirmed and contested parts are worth separating:
- Confirmed: The heap overflow has been reproduced independently on real hardware, and a remote request can crash the JVM. CISA describes the flaw as allowing unauthenticated remote code execution.
- Contested: Calif reports pre-authentication remote code execution in eleven Java products in their default configurations, but has not released its exploit code. Other analysts note that the overwritten bytes are fixed values rather than attacker-controlled, which makes code execution harder, though not impossible.
- Already circulating: VulnCheck says a full RCE exploit has been available to its commercial intelligence customers since the evening of November 17, and a public lab proof-of-concept has been posted.
Operational guidance: Treat the debate over exploitability as a reason to move quickly, not a reason to wait. A reliable crash is a denial-of-service bug on its own, and working exploit code is already in more than one set of hands.
Who's Actually Exposed
Unlike most "everything running Java" scares, this one has clear boundaries, and checking them is the fastest way to shrink the work. A system is exposed only when all three of these are true:
-
The Java version is affected. OpenJDK 17.0.6 and later, and every release from 18 up, including 21, 25, and 27, on a HotSpot JVM. That covers Oracle, Eclipse Temurin, Amazon Corretto, Azul Zulu, Microsoft, Red Hat, and Linux distribution builds. Java 8 and 11 are not affected, with one exception: Oracle's Java 8 Enterprise Performance Pack inherits the vulnerable code and should be treated like Java 17.
-
The CPU supports the accelerated path. The vulnerable routine runs only on x86-64 processors with the AVX-512 F, BW, VBMI, and BMI2 instructions, including Intel Xeon Ice Lake and newer and AMD EPYC and Ryzen Zen 4 and newer. ARM64 is not affected. Check from inside the workload, since a container or VM may expose a different instruction set than you expect:
grep -m1 '^flags' /proc/cpuinfo | tr ' ' '\n' | grep -E '^(avx512f|avx512bw|avx512vbmi|bmi2)$'If all four flags print, the hardware condition is met.
-
Attacker-controlled data reaches the Base64 decoder. This is more common than it sounds. HTTP Basic authentication decodes a header before checking credentials, which is why Tomcat and Jetty's Basic authenticators are named as pre-authentication paths, along with Spring Boot applications that use Spring Security's HTTP Basic support. JWT libraries such as Auth0's java-jwt decode tokens the same way. Reporting also flags Java clients that make outbound TLS connections to servers an attacker controls.
Watch for bundled Java. CISA notes that JDKs and Java runtimes may not show up as separate entries in software inventories. Elasticsearch, for example, ships its own JDK, so patching the host's Java does not fix it. Appliances, vendor products, and container base images carry Java their owners rarely think about.
What to Do
- Inventory by the three conditions. Find Java 17+ runtimes on x86-64 hosts with AVX-512 VBMI, then prioritize the ones that are internet-facing and decode untrusted Base64, starting with anything behind HTTP Basic authentication or handling JWTs.
- Patch the JDK. Oracle's fixed builds are 17.0.21.1, 21.0.13.1, 25.0.5.1, and 27.0.1.1. Other distributions ship the fix under their own version numbers and on their own schedules, and some Linux packages had not shipped rebuilt versions as of this writing, so verify against your vendor's advisory rather than a version number alone. No application or framework update fixes this; the fix lives in the JVM. Restart Java services after patching so the fixed runtime is actually running.
- Use the stopgap if you can't patch yet. Starting the JVM with
-XX:-UseBASE64Intrinsicsdisables the vulnerable accelerated path and nothing else, at some cost to Base64 performance. Remove it once you are on a patched build. Do not rely on proxy request-body buffering as protection; WebSocket traffic can bypass it. - Move off non-LTS Java. Java 26 and other non-long-term-support releases have reached end of updates and will not receive a fix. Move to 21 or 25.
- Apply CISA's compensating controls where patching lags: validate and size-limit Base64 input at the application boundary, and restrict unauthenticated access to exposed Java services.
- Monitor. Watch for unexpected JVM crashes (especially repeated
hs_err_pidcrash logs), unfamiliar Java runtimes, and suspicious child processes spawned by Java services. A spike in crashes on an internet-facing Java service is worth treating as possible exploitation attempts.
Summary Status
| CVE | Product | Fixed In (Oracle) | KEV Status | |---|---|---|---| | CVE-2026-114938 | HotSpot JVM (Java 17.0.6+, 21, 25, 27; x86-64 with AVX-512 VBMI) | 17.0.21.1 / 21.0.13.1 / 25.0.5.1 / 27.0.1.1 | Not yet listed | | Java 26 and other non-LTS | End of updates | No fix coming | Upgrade to 21 or 25 | | Java 8 and 11 (standard builds) | Not affected | — | — |
There is no BOD 26-04 deadline yet. If CISA adds this CVE to the catalog, we will publish a follow-up with the dates and what changed since this edition.
Sources: CISA Alert: Critical Java Vulnerability Requires Immediate Patching · Oracle Critical Security Patch Update, November 2026 · HeroDevs: Java4Shell in Spring Boot, Tomcat, Jetty · Endor Labs: Pre-Auth Heap Overflow in OpenJDK · Cyble: Java4Shell, Who's Exposed and How to Fix It · VulnCheck: Java4Shell · ComputingForGeeks: Find and Patch Vulnerable JVMs · Cycode: Java4Shell · CISA KEV Catalog
Free KEV Alerts
- Real-time notification the moment a KEV drops
- Vendor and product details
- BOD 26-04 remediation deadline included
Pro Alerts Coming Soon
- Real-time notification the moment a KEV drops
- Filtered to your specific vendor watchlist
- Urgency scoring (Critical / Urgent / Standard)
- Direct patch links included
Stay ahead of CISA.
Search the KEV Catalog by Vendor or Product
Search for CVEs by vendor or product to identify known exploited vulnerabilities in your environment
Upcoming Patch Due Dates
via Binding Operational Directive 26-04
BOD 26-04 is CISA's current vulnerability remediation directive for Federal Civilian Executive Branch (FCEB) agencies, updating the KEV-driven framework introduced under BOD 22-01 with a more risk-based approach to prioritization. While binding only on FCEB agencies, its framework increasingly influences contractor expectations through procurement requirements, FedRAMP programs, and agency security clauses.
Loading...
Cyber Security News
You may have missed...
Hacking Editorial Brief — November 18, 2026
Supply Chain Attack Compromises Major Websites via Optimizely Platform
Attackers compromised Optimizely's Web Experimentation platform on November 16, injecting malicious JavaScript into at least 178 organizations' websites, according to Bleeping Computer and The Register. Affected sites include BBC, CNN, Fox News, Zoom, HSBC, Salesforce, and GitLab. Researchers report the malicious code targeted visitors' cryptocurrency wallets, credit card data, and iPhone devices. The attack vector exploited a flaw in Optimizely's platform to load card skimmers and malware across customer sites, according to CXM World. The incident represents one of the largest recent supply chain attacks by affected site count, though Optimizely has not publicly disclosed the specific vulnerability, compromise timeline, or whether customer notification has been completed.
Critical Java Vulnerability Enables Pre-Authentication RCE Across Enterprise Platforms
Oracle and OpenJDK released patches for CVE-2026-114938, a Java flaw researchers have designated "Java4Shell" that enables unauthenticated remote code execution on widely deployed applications including Jira, Jenkins, and Keycloak running on AVX-512-enabled Java servers, according to The Hacker News. Separately, attackers are actively exploiting CVE-2026-119618, a critical race condition in Apache HTTP Server's HTTP/2 module that allows unauthenticated remote code execution, according to GBHackers. Researchers also observed exploitation of CVE-2026-84869, a ScreenConnect vulnerability with CVSS 9.9, enabling attackers to execute malicious files on managed service provider technician workstations without user confirmation, according to Offseq Radar. Health-ISAC reports threat actors are actively exploiting a critical Atlassian Confluence RCE vulnerability, though specific CVE designation was not provided in available reporting.
Espionage and Cybercrime Operations Target Cloud Infrastructure and Developer Communities
Google reports Russian espionage actor Star Blizzard is conducting phishing operations as recently as November 2026, abusing cloud identities for mass data theft, according to Infosecurity Magazine. A campaign designated Trigram impersonates approximately 50 AI and blockchain brands to deliver infostealers targeting developers across Europe and the U.S., according to Western Illinois University citing The Hacker News. Microsoft warned of a coordinated phishing campaign exploiting Microsoft 365 service disruptions to deliver fake service notifications across multiple sectors worldwide, according to Cybersecurity News. Additionally, U.S. authorities extradited Kazakh national Roman Khlynovskiy from Italy on charges of participating in ransomware attacks against healthcare organizations, schools, and companies as part of the 8-digits Team Inc. and Lynx groups, according to Recorded Future. A British national received a six-year federal prison sentence for conspiracy to commit wire fraud and aggravated identity theft after hacking companies and stealing data, according to DOJ.
Sources: Bleeping Computer · The Register · CXM World · The Hacker News · GBHackers · Offseq Radar · [Health-ISAC](https://health-isac.org/health-isac-daily-cyber-
*
Inside a cyberattack: How hackers steal data
The truth about cybersecurity is that it's almost impossible to keep hackers outside of an organization, particularly as the cybercrime industry ...
Java4Shell Flaw (CVE-2026-114938) Enables Pre-Auth RCE in Jira, Jenkins, Keycloak and More
A Java flaw called Java4Shell can let unauthenticated attackers run code on several widely used applications on affected AVX-512 Java servers; Oracle ...
Optimizely hack infected CNN, Fox News, BBC, Salesforce sites with malware
A suspected supply chain attack targeting Optimizely has led to malicious JavaScript being loaded on numerous popular websites, including Zoom, ...
Google Finds Hackers Using AI Agents to Build and Run Exploits From Software Patches
Google's threat intelligence group reports threat actors deploying autonomous multi-agent offensive frameworks to run multi-stage intrusions at machin...
Chinese state group developing automated attack tool for cyber espionage, Google says
Google's AI Threat Tracker report says a Chinese government-linked group is building its own version of the Artex AI hacking framework for likely use ...
Optimizely hack turns world's biggest websites into malware dispensers - The Register
BBC, Zoom, HSBC, and GitLab among those caught up in supply chain attack targeting visitors' crypto wallets, credit cards, and iPhones.
Google says Russian-speaking hackers ran the first fully automated AI cyberattack
Google's Threat Intelligence Group says the tools trade stealth for speed, leaving noisy forensic traces even as they succeed on their own.
China AI Espionage — US Warns of Total Mobilization for Tech Theft
US intelligence warns China is mobilizing government, private firms, and citizens using honey traps and hacking to steal advanced AI and semiconductor...
Google Reveals How Hackers Are Using AI Agents to Automate Real-World Cyberattacks
Google's Threat Intelligence Group reports that threat actors are actively using open-source AI-agent hacking frameworks like ARTEX, CAIRN, and PentAG...
Updated daily
