IT Security News Roundup: 2026-11-19 Overnight

IT Security News: Overnight roundup, 2026-11-19

Summary

The dominant story in this overnight window was CVE-2026-114938, a flaw in OpenJDK's Base64 decoder that Calif.io nicknamed Java4Shell. It drew urgent patch warnings from CISA, AWS and researchers, and an exploit attempt disguised as Chrome traffic showed that attackers were already moving. Other stories fit the same pattern: active exploitation of a Puppet flaw, malicious JavaScript injected through Optimizely, and AI appearing as both a new attack surface and a policy question. The practical takeaway for defenders is that patch priority should follow evidence of exploitation and exposure, not severity scores alone.

  1. IBM's Cost of a Data Breach Report 2026 says incidents involving employee shadow AI more than doubled in a year, from 20% to 43%. The piece argues the risk goes beyond staff pasting company data into unapproved chatbots.
  2. WeLiveSecurity explains how "wangiri" one-ring phone scams work, where a single missed call is designed to prompt a callback. It also covers how to keep these scammers at bay.
  3. The Rust project highlights four major changes for the language as of late 2026, including a next-generation trait solver designed to handle self-referential types and async features. The list also includes a new borrow checker called Polonius.
  4. eSecurityPlanet compares Fortinet FortiGate and Palo Alto Networks next-generation firewalls for 2026, covering security features, pricing, management capabilities and pros and cons for each.
  5. Risky Bulletin's hosts discuss a Trump administration AI task force asking US regulators to go easy on organisations hit by AI-powered attacks. They see this as a short-term step and call for policy that encourages defensive AI experimentation, alongside tighter enforcement of AI chip export rules.
  6. eSecurityPlanet reports that CVE disclosures passed 100,000 in 2026 and argues that severity scores alone are not enough to set patch priorities. Teams should also weigh exploitation data, asset exposure and business impact.
  7. The PHP security team warned that its updates, rated critical for the highest-severity issues across all supported branches, are due on 19 November 2026 between 14:00 and 14:30 UTC. Users are urged to prepare and install them promptly.
  8. AWS published a bulletin on CVE-2026-114938, a memory issue in the OpenJDK Base64 decoder intrinsic on x86-64 processors with AVX-512 VBMI. A fix is available in the November 2026 OpenJDK updates, and AWS says it is deploying the fix to affected services.
  9. Palo Alto Networks agreed to buy AI startup Applied Compute in an all-stock deal valued at $4.5 billion. The company says the deal gives it a platform for fine-tuning and running open-weight AI models in public and sovereign clouds.
  10. Cloudflare's Cloudforce One reported a Java4Shell exploit attempt that looked like ordinary Google Chrome traffic. The source first scanned the target with a User-Agent that revealed the scanning tool, then about 17 hours later sent the exploit attempt.
  11. An opinion column argues that AI is increasingly solving problems people thought unsolvable, pointing to an OpenAI model that disproved a central conjecture in discrete geometry. The author expects AI-written code to be overtaken by agent-written work that humans will struggle to review.
  12. WordPress 7.1.5, released on 18 November 2026, fixes 20 security issues in core. They include a flaw that could let an unauthenticated visitor run code on a vulnerable server, two SQL injection flaws, and a weakness that could let a Multisite administrator take over the network.
  13. Calif.io disclosed CVE-2026-114938, rated 9.2 critical under CVSS v4.0, a heap memory-corruption flaw in OpenJDK's hardware-accelerated Base64 decoder. It demonstrated pre-authentication remote code execution in default configurations of products including Keycloak, Jira, Confluence and Jenkins.
  14. The UK National Audit Office claims government risk management is falling short, with failures leading to poor government decisions. The report was covered by Infosecurity Magazine.
  15. The SANS Internet Storm Center published its Stormcast for 19 November 2026. The indexed entry has no text preview, so its content is not summarised here.
  16. Microsoft announced Windows Server 2028, the next major release of its server operating system, at the Ignite 2026 conference, with a focus on hardened security. BleepingComputer reported the announcement.
  17. CISA urged organisations to patch CVE-2026-114938 in Oracle Java immediately, warning that the out-of-bounds write can lead to denial of service or remote code execution under certain conditions. The Hacker News reports its CVSS score as 8.1.
  18. A guest diary from an ISC intern argues that honeypot analysis should be treated like code, using Git, pull requests and a human gate when AI is involved. The indexed teaser gives few further details.
  19. Smashing Security episode 493 covers how changing just a handful of numbers can turn an AI model into a sleeper agent. It also features NordLayer's advice on seeing which AI tools staff really use and stopping sensitive data leaking without banning AI.
  20. A UK parliamentary committee wants an AI watchdog with the power to pull the plug on dangerous models. It argues tech giants cannot be left to mark their own homework on AI safety, according to The Register.
  21. Europol and law enforcement agencies from 14 countries ran Operation TornADO, identifying vendors and buyers on major dark web marketplaces. From May 2025 to October 2026 it led to 354 arrests and hundreds of house searches worldwide.
  22. CISA added the Perforce Puppet flaw CVE-2026-115731, rated 9.4, to its Known Exploited Vulnerabilities catalog after reports of active exploitation. The flaw affects the configuration management platform used on enterprise management servers.
  23. An attacker abused a signup flaw in Optimizely's Web Experimentation on 16 November 2026 to inject malicious JavaScript into customer projects. ThreatLabz found more than 450 domains serving the script, with a payload designed to swap crypto wallet addresses and skim card details.
  24. Europol's EU Internet Referral Unit, working with Spain, ran a Referral Action Day against violent extremist and terrorist live-action role-playing in online gaming environments. Specialised units from 16 countries took part.
  25. The US government officially approved the Secure SI by Design Consortium, an industry coalition led by Palo Alto Networks with NVIDIA, IBM, OpenAI, Snowflake, Cognizant, Reflection, Factory and ServiceNow. Palo Alto Networks announced the approval on its blog.

Summaries written with AI from the linked source articles.

25
articles summarized
17
sources

Sources in this roundup

Cyber Security News
3 article(s)
InfoWorld
2 article(s)
News
2 article(s)
SANS Internet Storm Center, InfoCON: green
2 article(s)
Security Research | Blog
2 article(s)
The Hacker News
2 article(s)
esecurityplanet
2 article(s)
BleepingComputer
1 article(s)
Blog
1 article(s)
GRAHAM CLULEY
1 article(s)
Latest Bulletins
1 article(s)
Palo Alto Networks Blog
1 article(s)
Risky Bulletin
1 article(s)
securityweek
1 article(s)
welivesecurity
1 article(s)
www.infosecurity-magazine.com
1 article(s)
www.theregister.com – Articles
1 article(s)

Most-mentioned keywords

flaw
4 mention(s)
rce
4 mention(s)
security
4 mention(s)
java
3 mention(s)
shell
3 mention(s)
alto
2 mention(s)
base
2 mention(s)
code
2 mention(s)

Sources

  1. Shadow AI Is Shipped to You, One Release Note at a Time
  2. Who was that trying to phone me just now? How one-ring phone scams work
  3. Four major changes coming to Rust
  4. Fortinet vs. Palo Alto Networks: NGFW Comparison for 2026
  5. Srsly Risky Biz: Regulatory relief for AI hacks is a good first step
  6. 2026 CVEs Pass 100,000: Why Severity Scores Aren’t Enough
  7. PHP Security Team Warns of Critical Vulnerabilities, Urges Users to Prepare for November 19 Updates
  8. CVE-2026-114938 – OpenJDK Base64 Decoder Intrinsic Out-of-Bounds Write Issue
  9. Palo Alto Networks to Buy AI Startup Applied Compute for $4.5 Billion
  10. Hackers Exploiting Java4Shell RCE Flaw Using ARTEX Disguise Themselves as Chrome
  11. AI agents will soon surpass us
  12. WordPress 7.1.5 Patches 20 Security Issues, Including Potential RCE and SQL Injection Flaws
  13. CVE-2026-114938: OpenJDK Base64 Decoder Vulnerable To Pre-Authentication RCE (Java4Shell)
  14. Watchdog Warns UK Government Is Falling Short on Risk Management
  15. ISC Stormcast For Thursday, November 19th, 2026 https://isc.sans.edu/podcastdetail/10186, (Thu, Nov 19th)
  16. Microsoft announces Windows Server 2028 with hardened security
  17. CISA Urges Immediate Patching of Java4Shell Flaw That Could Allow Remote Code Execution
  18. Treat Your Honeypot Analysis Like Code: Git, Pull Requests and a Human Gate for AI [Guest Diary], (Wed, Nov 18th)
  19. Smashing Security podcast #493: Why your AI model could be a sleeper agent
  20. UK MPs want AI watchdog with power to pull the plug on dangerous models
  21. 354 dark web vendors and buyers arrested in Europol-coordinated global crackdown
  22. Hackers Exploiting Critical Puppet RCE Flaw Affecting Enterprise Management Servers
  23. Attackers Exploit Optimizely Flaw To Inject Malicious JavaScript
  24. Not a game: Europol and partners target terrorist role-playing online
  25. Secure SI by Design Consortium Officially Approved by U.S. Government