Daily Summary of Cybersecurity · Final editionThursday 19 November 2026 · UTC
Cybersecurity News: 19 November 2026
Palo Alto buys Applied Compute as Ukraine arrests ransomware gang
Palo Alto Networks agreed to buy Applied Compute for $4.5 billion to secure open-weight enterprise AI, while Ukraine arrested four key members of a ransomware network that extorted at least 28 companies for at least $40.5 million. CISA nominee Mike LoSapio faced a brief Senate hearing after the agency lost a third of its staff and had its budget slashed.
VMware's patched vCenter zero-day, linked to a suspected Chinese group, and Java4Shell exploitation attempts added to a day of patching and exploitation alerts.
CISA nominee LoSapio vows to find gaps after staff and budget cuts
Mike LoSapio, Palantir's chief information security officer and Trump's nominee to lead CISA, told a Senate committee he will identify gaps created by large staff and funding cuts. CISA has been without a Senate-confirmed chief for almost two years and has lost one-third of its staff, with its budget slashed by millions. Democrats said he struck a combative tone and took sides in partisan disputes, which could complicate confirmation.
VMware vCenter zero-day exploited by Chinese group before patch
Broadcom patched a critical VMware vCenter authentication bypass, CVE-2026-123890, that lets a network attacker gain an authenticated session. Mandiant linked the zero-day to Armory Castle, a suspected Chinese espionage group, which exploited it for more than two months after entering through a BeyondTrust flaw. The attackers used Bonesaw to steal database and SAML material and create an SSO administrator account, then installed backdoors.
PHP releases emergency patches for 18 critical vulnerabilities
The PHP Foundation released security updates for all supported versions to fix 18 vulnerabilities, seven of which are marked critical. The patches address flaws that could allow remote code execution, session corruption, and denial of service, including CVE-2026-126714, dubbed 'phpAfterShell,' which allows remote attackers to run arbitrary code by manipulating session files.
The release comes less than a month after the 'PHP2Shell' vulnerabilities triggered a global emergency patching scramble.
Chargemap breach exposes data of 3.5 million European EV charging users
Hackers breached Chargemap, an EV charging platform with over 3.5 million European users, stealing names, emails, phone numbers and birth dates. The company said the attack occurred on November 8, was discovered on November 10, and was patched on November 12, with the incident reported to France's CNIL. Customers were told on November 18 that no banking or financial data was affected, but exposed password hashes raise account takeover risk.
F6 said hackers exploited CVE-2026-76466, a critical Cisco Expressway flaw, to breach a Russian organization after public PoC code appeared. The flaw allows unauthenticated attackers with network access to clustered Expressway Core or Edge devices to run commands as root via persistent cookies. Cisco disclosed it on 21 October, fixed it in X15.5.20, and said there are no workarounds, though its advisory still says it is unaware of malicious use.
Drupal patches critical installer flaw that can allow unauthenticated RCE
Drupal patched a critical remote code execution flaw, CVE-2026-132165, in its core installer that can let anonymous attackers reinstall a site under specific conditions. Fixed versions are 10.6.22, 11.3.22 and 11.4.13, and the team said exploits may appear within hours or days. No attacks were reported, but sites should block /core/install.php or update, while Drupal Steward users are protected from known vectors.
n0n leaks Visa premium cardholder data after Visa confirms cyber incident
The n0n ransomware group published data it says it stole from Visa, claiming more than 1.5 million cardholder interactions tied to loyalty programs. Visa confirmed a cyber incident involving a third party and said no card numbers or payment data were affected. Cybernews found more than 6,000 premium Visa customers in the records, which point to dining and lifestyle campaigns in Asia-Pacific and could aid targeted phishing.
Apache confirms active exploitation of critical Roller and Axis flaws
Apache confirmed that attackers are actively exploiting CVE-2026-129448, a critical remote code execution flaw in Apache Roller versions 4.0.0 through 6.1.4. The vulnerability stems from unsafe deserialization in Roller's XML-RPC interface and allows unauthenticated attackers to execute arbitrary code. Apache also reported exploitation of CVE-2026-90436 in Apache Axis 1.x, which CISA added to its Known Exploited Vulnerabilities catalog.
Organizations are urged to upgrade to Roller 6.1.6 or apply temporary mitigations.
European Commission confirms AI-driven cyberattack that took systems offline
European Commission confirmed a cyber campaign detected October 22 and confirmed as an attack October 26, describing it as AI-generated and involving a publicly available tool. CERT-EU identified use of ARTEX, an open-source AI hacking platform. HR, payroll, medical reimbursement and translation systems were offline for about 10 days.
The Commission said affected teams were notified and it found no evidence of a staff data breach or ransom demand.
The n0n ransomware group posted data it says it stole from Walmart, claiming a staff directory, Sam's Club storefront maps, single sign-on settings and product metrics. Researchers found a smaller set with names and email addresses of about 78 employees plus product review counts, and no evidence that Sam's Club member data was accessed. The group's claims remain unverified, but the data could support phishing or credential-stuffing attacks.
TfL takes Oyster photocard site offline after account breach
TfL said about 265,000 Oyster photocard customers, including children, may have had account details compromised after emails from its systems asked them to create accounts they already had. TfL took the photocard website offline on November 16, contained the incident and notified the Information Commissioner's Office.
It said no passwords, bank or payment card details, or journey information were taken, and customers are asked to change passwords.
Ransomware group leaks 150,000 Samsung smart TV viewing records
The n0n ransomware group published 220MB of data it says it stole from Samsung after a leak timer expired. Cybernews found about 150,000 smart TV viewing events from 2024, including timestamps, durations, device details, cities, ZIP codes and provider names, but not what users watched. The files also included about 99,100 AI ad training records and two internal documents with employee details. The data could show political advertising spending.
SonicWall urges customers to replace Gen 6 firewalls now
SonicWall advised customers to replace Gen 6 firewalls immediately after all models running SonicOS 6.x reached end of support. The company called the fleet a high-risk exposure, noting that edge devices were targeted in 22% of recent breaches. Interim steps include disabling unused WAN interfaces and enforcing multi-factor authentication until migration to supported hardware is complete.
Splunk patches nearly 140 flaws in Enterprise, Security and SOAR
Splunk released patches for nearly 140 vulnerabilities across Enterprise, Enterprise Security and SOAR, including critical remote code execution flaws in Enterprise Security and SOAR. Enterprise fixes address two critical bugs that can let unauthenticated attackers read or modify clustering configuration or run arbitrary Search Processing Language searches. Splunk made no mention of any flaws being exploited in the wild.
nOn leaks JPMorgan data tied to Reserve card customers
nOn published data it says it stole from JPMorgan Chase after a 48-hour ransom deadline passed. The 502KB leak includes about 16,000 customer email addresses, some phone numbers, account IDs and 124 employee email addresses, all tied to the Reserve card for high-net-worth clients. JPMorgan said a service outage was caused by a software bug, not a cyberattack, but has not commented on the leak.
Researchers said the data could enable targeted phishing.
n0n ransomware leaks small alleged Mastercard customer dataset
The n0n group published a 280KB dataset it says was stolen from Mastercard after a 48-hour ransom deadline expired. It contains about 2,000 personal data entries, but Cybernews researchers said the records are poorly formatted and may be test data. The group had threatened to release cardmember profiles, bank and BIN data, campaign histories and premium cardmember travel and dining bookings. Its origin and authenticity are unverified.
N-able says hackers breached Mail Assure server, disabling management tools
N-able said unknown attackers accessed one server in its Mail Assure and SpamExperts Hosted Cloud, disabling management interfaces and APIs. The company said core mail flow and filtering remain available, but the management portal, APIs, quarantine and logging are offline while restoration is underway. It has no evidence customer data, including tenant mailboxes, was compromised, and said the incident is unrelated to recent N-central flaws.
NVIDIA says no sign of breach as Gentlemen ransomware posts countdown
The Gentlemen ransomware group added NVIDIA to its dark web leak site and set a countdown ending November 28, but has not yet published proof. NVIDIA said it found no indication its systems or networks were affected and is investigating an incident that may have involved a third-party partner. The group, which operates as a double-extortion RaaS, has claimed more than 1,100 victims since emerging last year.
F5 said it patched 22 vulnerabilities, including five critical NGINX flaws dubbed NGINX2Root that can allow unauthenticated RCE. Xint and other researchers found the bugs in HTTP/2, HTTP/3, mail and proxy configurations, with one affecting versions 1.9.5 through 1.31.8 and fixed in 1.31.9. Xint chained the worst flaw with a Linux kernel bug to reach root, and Censys says nearly 10 million Internet services run vulnerable versions.
Rakuten Securities breach could affect nearly 16 million customers
Rakuten Securities said a test environment breach may have affected nearly 16 million accounts, with 4.47 million exposed to names, dates of birth, addresses, employer, withdrawal account details and asset information. A further 11.25 million accounts had branch/customer codes and login IDs exposed. The firm said passwords and PINs were not taken and warned customers of phishing, urging those not using passkeys to switch.
Java4Shell exploitation attempts begin as vendors rush advisories
Attackers have begun probing for Java4Shell, CVE-2026-114938, a memory corruption flaw in OpenJDK 17.0.6 and newer that can allow unauthenticated remote code execution on x86-64 systems with AVX-512 VBMI. Calif disclosed it on November 17, the same day Oracle and OpenJDK released patches. GreyNoise recorded over 150 matching requests from five IP addresses, while WatchTowr and Cloudflare reported reconnaissance and early exploitation attempts.
Ruby fixes seven regex and string flaws, five with RCE PoCs
Ruby maintainers released versions 4.0.10, 3.4.13 and 3.3.14 to fix seven memory safety flaws in its regular expression engine and string case conversion code. They said five have proof of concept exploits that can run code, while the remaining two are weaker, allowing memory corruption, improper reads or crashes. The advisories rate three flaws critical, three high and one moderate, and report no attacks so far.
US Threat Disruption Center enlists private firms to stop scam payments
The Homeland Security Task Force launched the Threat Disruption Center to sort complaints, map criminal groups and connect investigators with private firms. Firms can pause payments, take down malicious domains and ban scammer social media accounts under their terms of service, while agencies handle investigations, prosecutions and seizures. It builds on Operation Level Up, after $20.9 billion in losses were reported to the FBI's IC3 in 2025.
Researcher shows database RCE exploits in PostgreSQL, MySQL, Oracle, ClickHouse
Depthfirst researcher Zheng Yu published proof of concept exploits showing low privilege database accounts can run code on the server by reaching memory corruption bugs in PostgreSQL, MySQL and ClickHouse, while a separate Oracle chain needs no login. Four flaws now have vendor CVEs and fixes, but full attack chains were not independently reproduced and no real attacks have been seen.
CISA adds four exploited flaws tied to ransomware, cryptomining, LLMjacking
CISA added four actively exploited flaws to its Known Exploited Vulnerabilities catalog, affecting Synology DSM, BerriAI LiteLLM, Apache Axis and Vanderbilt REDCap. One Axis flaw, CVE-2026-90436, was linked to cryptocurrency miners and was seen in Atlassian Confluence Data Center, where attackers ran scripts to hunt credentials, tokens, VPN settings and payment card data. The incident was flagged on November 13.
Game consoles and Teslas slip past OT network segmentation
Claroty found tens of thousands of consumer devices, including game consoles, Roku sticks and Teslas, on hospital, factory and commercial networks after analyzing over 45 million devices at over 30,000 critical infrastructure sites. Game consoles mostly sat on production networks, while cars and Sonos speakers mostly stayed off guest networks. A compromised device could let attackers look for controllers, medical devices and building systems.
AI agents inherit employee access in company tools
Nudge Security found that AI agents built in project trackers, CRMs and automation tools often run on the credentials of the employee who created them, carrying that person's access to company data. The report says the median organization with at least one agent had seven, and more than half of organizations with agents ran them on two or more platforms, complicating audit trails.