Administrators running sshd on Alpine Linux servers are being told to update openssh and restart sshd right away, since upgrading the package alone leaves the old daemon running. A flaw in the OpenSSH server lets attackers log in without a password. Those who cannot patch yet can close the hole by setting PasswordAuthentication to no and restarting sshd. 

How the OpenSSH Server Flaw Works 

An attacker who can reach the SSH port gets in with a single connection that works every time, with no retries. The attacker needs the name of an account whose password field is * (normal for key-only accounts) plus at least one other account with a $5$ or $6$ hash, then sends a password of 257 to 1,024 bytes. One user with a password exposes every account without one.  When built without PAM, sshd hashes each attempt with crypt() and compares it to the stored field, borrowing another account's salt when that field is too short to hold one. musl's crypt() returns * as an error indicator for keys over 256 bytes, which exactly matches the stored field of passwordless accounts. OpenSSH 10.6p2 fixes this, and no CVE had been assigned at the time of writing. 

Which Alpine Linux Servers Are Affected 

Alpine Linux servers are vulnerable when they run openssh-server (built without PAM) or openssh-server-pam with UsePAM no, keep the default PasswordAuthentication yes, hold at least one local $5$ or $6$ hash, and have an account with a * password field. That includes root, shipped as root:* in the base image since about Alpine 3.16. The default PermitRootLogin prohibit-password shields root unless it is set to yes, but not if another account can become root through doas. Key-only users created with adduser or usermod -p '*', directory users via musl-nscd (nslcd, winbind, sssd), and the tiny-cloud alpine user also count. Every release checked is affected: 3.21 through 3.24, edge, and the older 3.16 to 3.20.  Safe setups include containers not running sshd, systems with PasswordAuthentication no or UsePAM yes, locked accounts (user:! in /etc/shadow), machines with no $-style hash, and cloud-init images, which write PasswordAuthentication no. 

Patching the OpenSSH Server 

Running apk upgrade -U -a, followed by rc-service sshd restart, applies the fix, which ships in Alpine 3.21.10, 3.22.8, 3.23.8 and 3.24.4. Fixed builds are edge 10.6_p2-r0, 3.24 10.3_p1-r2, 3.23 10.2_p1-r1, 3.22 10.0_p1-r11, 3.21 9.9_p2-r1, 3.20 9.7_p1-r6, 3.19 9.6_p1-r3, 3.18 9.3_p2-r4, 3.17 9.1_p1-r7 and 3.16 9.0_p1-r6. Although 3.21 reached end of support on 2026-11-01, patched packages go back to 3.16.  For the workaround, root should first confirm key login works and keep the current session open. A drop-in file, /etc/ssh/sshd_config.d/00-no-passwords.conf (read from 3.19 onward), sets PasswordAuthentication no, while a sed command flips every such line in sshd_config, Match blocks included, and saves sshd_config.bak. Running sshd -t && rc-service sshd restart, then sshd -T | grep -i passwordauthentication, should print passwordauthentication no. Deleting the drop-in and restoring the backup undoes it. 

Cloud Images and Spotting Attacks on Alpine Linux Servers 

Official tiny-cloud images (qcow2, VHD and AMI) from 3.18.11 through 3.24.3 ship an alpine user with a * field and passwordless doas. They are safe on first boot but become exploitable once any $6$ hash appears, such as a console password or a user-data account with hashed_passwd, giving attackers a shell as alpine and then root. Instances from earlier point releases need updating or password logins disabled.  Admins should check /var/log/messages for sshd "Accepted password" lines on passwordless accounts and review doas use, new users, files, and authorized keys. If a login cannot be ruled out, the machine should be treated as compromised.  Özgün Kültekin of Wakefield reported the issue. Damien Miller and Darren Tucker of the OpenSSH project delivered the fix and the 10.6p2 release; Timo Teräs helped analyze the bug, and Natanael Copa built and uploaded the updated packages. 

Attackers needed less than two days to start going after a freshly patched WordPress vulnerability. The flaw sits in WordPress core itself, and it can let an attacker run code on a website while that site is installing an update.  WordPress shipped version 7.1.4 on October 28, 2026, to close the hole. Within 48 hours of that release, exploitation attempts were already underway. Security firm Wordfence said it blocked 62,969 attacks aimed at this vulnerability in WordPress over a single 24-hour period.  Anyone running WordPress for a business, a nonprofit, or a healthcare facility is being urged to update right away. 

What WordPress 7.1.4 Fixes Beyond the WordPress Vulnerability 

Version 7.1.4, released on October 28, is mainly a security release. Alongside the critical code execution fix, it patches several stored cross-site scripting (XSS) flaws, a denial-of-service issue, and two separate information leaks.  The update also refreshes WordPress's bundled root certificates after eMudhra revoked its certificates. Of everything in the release, the code execution bug is the one attackers zeroed in on.  Aaron Jorbin, who wrote the release announcement, told users to “update your sites without delay.”  A GitHub security advisory published the next day described the issue as a timing weakness in the update system. If an unauthorized user moves quickly enough while an upgrade is running, they may be able to execute their own code on the server. No CVE number has been assigned to it so far. 

How the Vulnerability in WordPress Works 

When WordPress performs a full-package core update, it first extracts the new files into a dedicated directory inside the upgrade folder. According to the advisory, that directory had a predictable name and was reachable from the public web.  That combination is the problem. An attacker who requests the right file at the right moment, before WordPress finishes cleaning up, could trick the server into running arbitrary PHP code.  Pulling this off is not trivial. The attacker needs four conditions to line up: a full-package update in progress, web access to the temporary folder, a path they can guess, and a short window before the files are deleted. Under controlled conditions, that window may last only a few seconds. 

Which Sites the WordPress Vulnerability Affects 

The advisory lists every WordPress release from 4.7 through 7.1.3 as vulnerable. It also names the fixed version for each branch, including 7.0.8, 6.9.11, 6.8.12 and 5.9.20.  The WordPress project said it backported the patch to all branches still receiving security maintenance, reaching as far back as 4.7. It also stressed that only the newest version is actively supported.  Not every site can be taken over through this bug. The risk rises sharply, however, if an attacker hits a site while it is running a full-package update.  Smaller websites face extra exposure. Many sit on shared servers with default settings, which makes them attractive targets for cybercriminals. 

How Attackers Are Targeting the Vulnerability in WordPress 

Wordfence traced the root cause to full-package upgrades repeatedly using web-accessible paths that are easy to guess. Alex Thomas, a threat analyst at Wordfence, said attackers quickly understood the race condition and turned their attention to it.  Two of the biggest attack sources reportedly kept polling small groups of websites for roughly 14.5 hours. Other sources tried out different folder names and version names.  The pattern looks like large-scale automated guessing rather than attacks on chosen victims. So far, no confirmed reports have put a number on how many sites have actually been compromised. 

How Serious Is This WordPress Vulnerability? 

Sources disagree on how severe the flaw is. The Hacker News gave it a CVSS score of 8.1, which falls in the high range. The GitHub advisory, by contrast, rates it as low severity.  That gap most likely comes down to how hard the bug is to exploit successfully. Even so, the sheer volume of attack attempts and the simplicity of the fix make it sensible to treat this as a serious threat. 

What Comes Next for This Vulnerability in WordPress 

The probing is unlikely to stop soon. Once a vulnerability is patched, the fix itself becomes a detailed roadmap for anyone hunting sites that have not been updated.  The WordPress project says backports to older branches are still in progress. The next major release, WordPress 7.2, is scheduled for December 8.  The advisory credits researcher Umut Bayram with discovering the flaw and reporting it responsibly. There is currently no alternative workaround, so applying the patch is the only fix available. 

A decade-old flaw in the language behind much of the web, found by AI and serious enough to take Wikipedia offline. Driving licenses sold for AED 10,000 to people who never sat for a test. A police sweep across 21 African countries that ended in more than 1,400 arrests. 

This weekly roundup covers stories where the weak point was not always a firewall, but the trust built into code libraries, license agreements, and the frameworks organizations use to measure their own security. 

Here are the stories that shaped the week. 

The Cyber Express Weekly Roundup 

PHP2Shell Flaws Knock Wikipedia and Government Sites Offline 

Two critical vulnerabilities in PHP, the scripting language that powers a large share of the web, have triggered urgent patching warnings from agencies around the world. Dubbed PHP2Shell and tracked as CVE-2026-109618 and CVE-2026-109619, the flaws have already been linked to outages at Wikipedia and a number of government websites. Read more… 

INTERPOL’s Operation Serengeti III Nets 1,462 Arrests Across Africa 

An INTERPOL-coordinated crackdown spanning 21 countries has ended in 1,462 arrests. Operation Serengeti III ran from June to August 2026 and targeted fraud, ransomware, business email compromise and online scams. Investigators disrupted 10,909 malicious networks and recovered more than USD 105 million. The operation also identified nearly 66,000 victims, linked to combined losses of more than USD 385 million. Read more… 

ASD Essentials Replaces the Essential Eight 

The Australian Signals Directorate has launched ASD Essentials, a broader successor to the Essential Eight, the security baseline Australian organizations have long been measured against. Where the Essential Eight centered on a fixed set of mitigation strategies, the new framework extends its scope to enterprise IT, operational technology and cloud environments. Read more… 

Dubai Police Bust AED 10,000 Fake Driving License Scheme 

Dubai Police have arrested an eight-member gang that sold driving licenses for AED 10,000 each. Customers received licenses issued in a neighboring country without ever travelling there, taking a lesson or sitting a test. The gang then helped buyers convert those licenses into UAE licenses by exploiting legitimate license exchange agreements between the two countries. Read more… 

Fintechs Urged to Treat Software Supply Chain Security as Enterprise Risk 

Fintech firms vet their commercial vendors closely, but the open-source libraries, container images and third-party components inside their systems often arrive with far less scrutiny. That mismatch leaves a blind spot in how financial technology companies manage risk. Vijendra Katiyar of CleanStart points to examples such as a widely used Redis image that carried 185 known CVEs, and the Shai-Hulud worm, which spread through malicious npm packages. Read more… 

Conclusion 

Taken together, this week’s stories show how much of today’s risk comes from what organizations already trust: decade-old code, inherited software components, official exchange agreements and security checklists that measure effort rather than results. Follow The Cyber Express for next week’s weekly cybersecurity roundup and ongoing coverage of these stories. 

site-verification-ok

site-verification-ok

Japan is preparing to set up a dedicated team to run AI-driven cyberattack simulations to expose weaknesses in technology systems. Japan’s plan for AI cyberattack simulations comes as organizations worldwide scramble to address critical PHP flaws that have disrupted parts of the internet. 

Speaking on Friday, Industry Minister Ryosei Akazawa said those defending against cyber threats need to think like attackers, using AI proactively to uncover potential weak points before they can be exploited. 

His comments follow widespread disruption caused by the disclosure of the PHP flaws, which led to millions of reported hacking attempts. Businesses and governments were forced to patch their systems or temporarily take websites offline.  

Because PHP is widely used to build and run websites and web applications, any weakness in the language poses a risk to organizations across many sectors. 

How PHP Flaws Hit Banks, Airlines and Government 

The vulnerabilities triggered precautionary shutdowns across banking, transport, airline and government services in several countries. 

In Japan, Nomura, Mitsubishi UFJ Financial Group, ANA and East Japan Railway temporarily halted some online services. The Ministry of Finance, the National Tax Agency and the Meteorological Agency also suspended certain digital services as part of their security response. 

On Friday, JR East said some of its services were back online, while others would resume only after safety checks. Nippon Paint said it had applied a security patch and restored its website. 

The disruption shows how difficult it is to secure interconnected digital infrastructure when a widely used software component turns out to have serious weaknesses. 

SoftBank Pauses Emergency Alerts Over PHP Flaws 

The fallout also reached services people rely on in a crisis. Telecommunications company SoftBank said it was temporarily suspending its mobile emergency alert system while it carried out maintenance to fix a PHP vulnerability. 

While the system is down, customers will not receive early warnings through it for earthquakes, tsunami threats, disasters, or evacuation information. The case shows that cybersecurity incidents can extend well beyond ordinary websites and interrupt access to vital public safety information. 

Why Japan AI Cyberattack Simulations Point to Proactive Defense 

The episode has put a fresh focus on finding vulnerabilities before attackers do. The Japan AI cyberattack simulations are designed to help defenders understand how weaknesses might be targeted and to sharpen their response capabilities, in line with the approach Ryosei Akazawa outlined. 

Simulation alone is not enough, however. Installing patches remains essential, and organizations need to test updated systems carefully before bringing services back, especially where transport, finance, and government operations are affected. 

As AI tools grow more capable for both attackers and defenders, businesses and governments face the task of pairing fast vulnerability detection with effective security updates, continuous monitoring, and resilient digital infrastructure. 

site-verification-ok

site-verification-ok

Governments and cyber-security agencies worldwide are urging website operators to patch PHP2Shell, two critical PHP flaws tracked as CVE-2026-109618 and CVE-2026-109619. Found with AI, the bugs enable remote code execution and Heartbleed-style memory disclosure. A fix arrived on Wednesday afternoon. 

Several agencies told organizations unable to patch in time to consider pulling their sites. Finland temporarily shut its government’s public web services. In the Netherlands, the cyber-security agency, the National Institute for Public Health and the Environment, and the Tweede Kamer, parliament’s lower house, all went offline. 

The US National Security Agency echoed this on X: “These vulnerabilities often provide a direct path to the server’s file system or database.” The White House said federal agencies were working with leading AI labs to speed up a fix, while Japan’s Prime Minister Sanae Takaichi urged people on X to check whether their public websites ran on PHP. 

“We have taken down a number of other systems as a precautionary measure until they can be safely patched,” the Wikimedia Foundation said, leaving parts of Wikipedia unreachable on Wednesday. Edinburgh council’s website and pages for BBC programmes, including Doctor Who and The Traitors, also went dark, though neither has confirmed a PHP link. DeviantArt said only that it was “temporarily down for maintenance.” 

How AI found CVE-2026-109618 and CVE-2026-109619 

Researchers at Calif and Wiz uncovered the flaws. Calif and the PHP Foundation were part of OpenAI’s Patch the Planet program, which offered AI tools to find and fix PHP flaws. Wiz called them the “most severe vulnerabilities ever discovered” by AI. OpenAI described them as “one of the most impactful security issues in recent memory,” adding that its own researcher spotted the issue around the same time. 

The bugs sat undetected for over a decade before surfacing August, after a few weeks of AI-assisted work. “Frontier models were critical to making this process practical, enabling us to develop and adapt working exploits across multiple applications in hours,” Wiz said. It estimates “roughly 70%” of the internet is affected; a figure not independently verified. 

The PHP Foundation planned to disclose PHP2Shell later this month with a full patch, but details leaked on social media on Tuesday, forcing a rushed announcement. 

Early PHP2Shell Attacks Emerging 

The UK’s National Cyber Security Centre (NCSC) and its EU and Australian counterparts urged PHP users to apply the fixes. “Where immediate remediation is not feasible, organizations should consider taking the system offline until it can be fully mitigated,” the NCSC said. Japan’s national cyber-security office and EU officials issued similar warnings. 

“PHP is everywhere. It quietly powers WordPress, Drupal, Joomla, and Magento sites, shared hosting platforms, control panels, and an awful lot of software that doesn’t even advertise that it runs PHP under the hood,” said security firm Huntress. 

Affected versions remain unclear, though CERT-EU said CVE-2026-109618 and CVE-2026-109619 could reach back to PHP 5, launched in 2004. The update covers only the latest versions; users on PHP 8.1 or older should upgrade. Built for web development, PHP runs server-side on a large share, though not all, of the internet. 

site-verification-ok

The FBI, CISA and international partners have issued a warning about SafePay ransomware, a group that has impacted at least 600 victims across multiple critical infrastructure sectors since August 2024. Published on October 20, 2026, the joint cybersecurity advisory describes SafePay as among the most active ransomware groups in 2026, with operations affecting organizations in North America, South America, Europe and Australia.

The advisory was released by the FBI, the Cybersecurity and Infrastructure Security Agency (CISA), the United States Secret Service, Germany’s Landeskriminalamt Rheinland-Pfalz, the United Kingdom’s National Crime Agency and the UK’s North West Regional Organised Crime Unit. It draws on investigations conducted as recently as August 2026.

How SafePay Ransomware Gains Initial Access

According to the agencies, SafePay actors break into networks by stealing or guessing passwords, running brute force attacks, using phishing and exploiting VPN access. The group uses IP addresses from the VPN provider Mullvad and phished account profiles to exploit two flaws in edge devices: CVE-2020-3259 and CVE-2023-20269.

The actors also use automated scripts to exploit three vulnerabilities in internet-exposed systems: CVE-2023-27532, CVE-2024-40711 and CVE-2024-40766.

Living Off the Land Inside Victim Networks

Once inside, SafePay actors establish persistence by gaining administrative privileges on existing accounts or creating new ones. They rely on built-in tools such as PsExec, Windows Remote Management and Remote Desktop Protocol to execute commands and move laterally.

To find valuable files, the group has used ShareFinder.ps1 and SharpShares.exe to enumerate network shares. Before moving laterally, the actors commonly disable security software. In some cases, they changed system users’ passwords, locking legitimate users out. The advisory also notes the use of a User Account Control bypass, token impersonation and the credential dumper Mimikatz.

Data Exfiltration Before Encryption

Data exfiltration is a central part of the group’s playbook. SafePay actors compress files using 7-Zip, WinRAR, FileZilla and RClone, then move stolen data out through WinSCP, RClone, OneDrive and the Backblaze command-line tool b2.exe. Remote desktop tools such as AnyDesk and TeamViewer have been installed to transfer ransomware onto victim systems.

To block recovery, the actors halt the Volume Shadow Copy Service and reset network attached storage holding backups to factory settings.

Files are then encrypted using a hybrid scheme combining the ChaCha20 stream cipher with RSA public-key cryptography. Encrypted files carry the .SafePay extension and, according to the advisory, cannot be recovered without the decryption key.

The group typically demands a portion of company revenue in cryptocurrency. A ransom note left on the desktop or root directory directs victims to a TOR chat portal, using a victim ID to log in. A representative also makes phone calls to pressure victims. If payment is refused, stolen data is posted to SafePay’s leak site. The ransomware may delete itself after encryption.

SafePay Ransomware Advisory: Recommended Defenses

To reduce the risk of ransomware attacks from SafePay, the agencies urge organizations to prioritize patching known exploited vulnerabilities, especially on firewalls and other internet-facing systems. They recommend requiring multifactor authentication for webmail, VPNs and accounts that access critical systems.

Other measures include passwords of 15 to 64 characters, account lockouts after failed logins, offline backups that are encrypted and immutable, network segmentation, least privilege access and time-based admin access. The advisory also recommends restricting command-line and scripting permissions.

The advisory lists IP addresses, file names, hashes and hostnames observed between November 2024 and August 2026. The agencies advise verifying these indicators before blocking, as some IP addresses may later be reassigned to benign services.

The authoring agencies do not encourage paying ransom, noting payment does not guarantee file recovery. Incidents can be reported to the FBI’s Internet Crime Complaint Center, a local FBI or Secret Service field office, or CISA. In the UK, victims can report through Report Fraud or their local police force.

site-verification-ok

SAP released its October 2026 security patch on Tuesday, October 13, delivering 22 new and updated security notes. Four of them address critical-severity vulnerabilities. The most severe is CVE-2026-89111, a maximum-rated flaw in the Cloud SDK for Python. Several of the bugs could allow attackers to invoke internal operations or execute code, which could lead to complete compromise of affected systems.  SAP's patch day summary breaks the release down into 16 new security notes, four GitHub security advisories, and two updates to previously released notes. The company recommends that customers visit its support portal and apply the patches as a priority. 

CVE-2026-89111 Puts Cloud SDK for Python Users at Risk 

CVE-2026-89111 has a CVSS score of 10.0. It is an OAuth2 server-side request forgery (SSRF) vulnerability affecting multi-tenant BTP applications that use Cloud SDK for Python versions earlier than 0.45.2. It is tracked under GitHub advisory GHSA-4mv3-jxxm-9cx2.  SAP describes the risk in its advisory: "An attacker who controls the tenant_subdomain parameter can redirect OAuth2 token requests to an attacker-controlled server, causing client credentials to be exfiltrated." 

Other Critical Fixes in the SAP Security Patch 

Note 3797111 resolves CVE-2026-89110 (CVSS 9.9). This is a broken access control issue in certain portal functionality of NetWeaver Visual Composer, version VC70RUNTIME 7.50.  Note 3805589 fixes CVE-2026-89107 (CVSS 9.9), a code injection bug in S/4HANA's Manage Substitution/Validation Rules. It affects S4CORE versions 105 through 109.  Note 3798788 covers CVE-2026-89108 (CVSS 9.1), a missing authorization check in the Configuration Service of SAP Manufacturing Integration and Intelligence (XMII 15.4 and 15.5). 

High-Severity Vulnerabilities 

Three notes are rated high. Note 3800517 patches CVE-2026-89113 (CVSS 8.8), an SQL injection flaw in Commerce Cloud's OOTB code (COM_CLOUD 2211-JDK21).  Note 3765424 addresses CVE-2026-58249 (CVSS 7.5), a memory corruption bug in the Internet Communication Manager of NetWeaver and Web Dispatcher. It affects a wide range of components, including HDB 2.00, KRNL64UC 7.53, XS_ADVANCED_RUNTIME 1.00, SAP_EXTENDED_APP_SERVICES 1, and CONTSERV 7.53 and 7.54. It also covers WEBDISP 7.53, 7.54, 7.77, 7.93 and 9.16, and KERNEL 7.53, 7.54, 7.77, 7.93, 9.16, 9.18, 9.19 and 9.20.  Note 3626303 fixes CVE-2026-89120 (CVSS 7.1), a missing authorization check in the ABAP Platform (KERNEL 7.93 and 9.16). 

Medium and Low-Severity Notes 

Cross-site scripting (XSS) makes up a large share of the remaining fixes: 

• CVE-2026-76972 (6.5) affects Manufacturing Integration and Intelligence (note 3798823). 
• CVE-2026-89117 (6.1) affects the Enterprise Portal, EP-ADMIN 7.50 (note 3779985). 
• CVE-2026-76975 (6.1) affects the SAPUI5 Support Assistant Overlay (note 3794572). 
• CVE-2026-76976 (5.4) affects NetWeaver AS ABAP applications based on SAP GUI for HTML (note 3671819). 
• CVE-2026-89106 (4.4) affects SAPUI5 (note 3799054). 

Two updated notes also deal with older issues. Note 3689080, first released in March 2026, covers CVE-2026-24316 (6.4), an SSRF flaw in NetWeaver AS for ABAP that spans SAP_BASIS 740 through 918. Note 3629325, from September 2025, addresses CVE-2025-42938 (6.1), an XSS bug in the NetWeaver ABAP Platform.  The Enterprise Portal receives two more fixes. Note 3790401 resolves an open redirect, CVE-2026-89122 (6.1), and note 3789825 resolves an information disclosure flaw, CVE-2026-89118 (5.3). Both affect EP-RUNTIME 7.50.  The other GitHub advisories cover three open-source components: 

• CVE-2026-89112 (6.5) is a broken access control flaw in CLA-Assistant versions up to 2.13.1 that exposes signers' PII. 
• CVE-2026-89116 (5.9) is an OS command injection bug in Yeoman-UI version 1.25.1 and earlier. 
• CVE-2026-89105 (4.2) is an improper authentication issue in the HybridTokenAuthenticator of the cloud-security-services-integration-library, affecting versions 3.6.6 up to, but not including, 4.1.1. 

Missing authorization checks were also fixed in two more products. CVE-2026-89121 (4.7) affects S/4HANA Linear Asset Management (S4CORE 109), and CVE-2026-89124 (4.3) affects Global Trade Services, edition for SAP HANA (GTSCORE 200 and 300).  The only low-rated note, 3767634, addresses CVE-2026-58242 (CVSS 3.5), an improper authorization check in Cloud Connector 2.0. 

Microsoft's October 2026 Patch Tuesday fixes 1,413 Microsoft CVEs. That is a new record, set only one month after September's previous high of 974, and it represents a jump of roughly 45%.   The release covers Windows, Microsoft 365 and Office, SharePoint Server, SQL Server, Exchange Server, Azure, Defender and developer tools. It also republishes one non-Microsoft CVE. Two of the fixes are zero-days: one is under active exploitation and the other has been publicly disclosed. 

Severity Breakdown for October 2026 Patch Tuesday 

Of the 1,408 CVEs with severity ratings, 78 are Critical, 1,329 are Important and one is Moderate. All 1,408 require customers to install an update. Sixty-five of the Critical flaws allow remote code execution (RCE). The other 13 involve elevation of privilege or spoofing.  Elevation of privilege leads with 673 CVEs. It is followed by remote code execution (315), information disclosure (252), denial of service (72), security feature bypass (40), tampering (29) and spoofing (27). The first two types make up about 70% of the release. That combination is dangerous because RCE gives an attacker initial access and privilege escalation turns that access into full control.  Windows accounts for 1,102 vulnerabilities across 39 updates. The rest break down as follows: Office (134 vulnerabilities, 11 updates), developer tools (61, 60), SharePoint Server (36, 1), other products (32, 25), SQL Server (22, 8), Azure (15, 20), Exchange Server (7, 8), Defender (4, 3) and Edge (0, 1). Together that comes to 1,414 vulnerabilities and 176 updates. Microsoft's own figures don't fully match. The release notes cite 1,413, the product breakdown totals 1,414 and the published CVE list contains 1,411. This article uses the release-notes figure. 

Two Zero-Days, Both Rated Important 

CVE-2026-96951 is a security feature bypass in Windows Shell, and attackers are actively exploiting it. It defeats Mark of the Web, which means malicious files arriving by email or chat can open without the usual warnings. Users still have to open the file, so the flaw works well alongside phishing. It should be patched on every Windows endpoint first. CVE-2026-108398 affects the .NET Framework. It has been publicly disclosed but has not been exploited. It could let an authenticated attacker bypass Windows Defender Application Control and run code inside a protected process. That makes it a concern for organizations that rely on allowlisting to contain intruders. Both flaws are rated only Important, which shows that exploit status should matter more than severity labels. Neither appeared in CISA's Known Exploited Vulnerabilities catalog at publication, although additions often follow within days.

Critical Flaws in Network, Identity and SharePoint Server 

Windows DHCP Server has five Critical RCE flaws: CVE-2026-96934, CVE-2026-88117, CVE-2026-88120, CVE-2026-88119 and CVE-2026-88171. The DHCP Client has one more (CVE-2026-70250). Several other network services also carry Critical RCEs: Microsoft QUIC (CVE-2026-105016, CVE-2026-100219), Remote Desktop Gateway (CVE-2026-70187), Message Queuing (CVE-2026-108460, CVE-2026-100202) and RMCAST (CVE-2026-69789). Systems reachable from outside the network should be patched early.  Identity services are affected too. Active Directory Domain Services (CVE-2026-70229) and NDES (CVE-2026-69802) have RCE flaws, and AD FS has two spoofing bugs (CVE-2026-108131, CVE-2026-96222). Exchange Server (CVE-2026-70571) and Windows Deployment Services (CVE-2026-88174) each have a Critical RCE. SharePoint Server has a Critical elevation-of-privilege flaw (CVE-2026-102866).  Hyper-V receives fixes for host-level RCE (CVE-2026-102214, CVE-2026-105733), which could put every guest VM at risk, and for privilege escalation (CVE-2026-108399). VBS Enclave has five Critical elevation-of-privilege flaws: CVE-2026-108471, CVE-2026-108130, CVE-2026-102211, CVE-2026-96914 and CVE-2026-88074. Four more affect the Windows Kernel, Graphics Kernel, ALPC and the WSL Linux kernel (CVE-2026-96913, CVE-2026-88073, CVE-2026-83917, CVE-2026-62840), and these could give an attacker SYSTEM or root access.  Office apps account for 34 Critical RCEs: 19 in Excel, 5 in Outlook, 4 in PowerPoint, 3 in Word and 3 in Visio. The Mark of the Web bypass makes these more dangerous. Another ten affect media and rendering components, including Media Foundation, GDI+, Codecs, Windows PDF, Text Shaping and Raw Image Extension (CVE-2026-96202, CVE-2026-96943). They can be triggered just by opening a crafted file, with no macros involved. The PostScript Printer Driver, which is often overlooked, has three more (CVE-2026-70153, CVE-2026-70147, CVE-2026-108467). 

The XRP Ledger (XRPL) has deployed an emergency patch for one of the most serious XRP vulnerabilities in its history, a flaw that could have allowed unlimited XRP to be created. By skipping the standard two-week validator vote, the fix has reignited debate over how decentralized the network really is, though developers and validators maintain the step was necessary and did not amount to centralized control.  According to Woofun AI, the bug could have pushed XRP's supply beyond its predefined cap. The episode unsettled community expectations around transparent protocol governance, exposing a clear tension between fast security responses and decentralization.   XRPL officials said the goal was to stop attackers from exploiting the flaw, but sidestepping established procedure has cast doubt on the network's claim of "full decentralization," a narrative that underpins XRP's non-security status. The incident is effectively a stress test of governance under extreme risk, and its outcome is likely to shape how the market judges trust in XRPL's underlying architecture. 

The XRP Vulnerabilities Patch Update

The patch went live on September 25, just three days after the related bug bounty report was filed, and took effect as soon as servers installed it. That speed was unprecedented in the system's decade-plus history and marked the first deliberate bypass of the standard amendment process.   In a Sunday statement on X, XRPL developers explained that the customary two-week validator vote would have made the patch's code public weeks before activation, creating serious cybersecurity risks.  The update was initially released only in binary form, with its source code temporarily withheld. Data compiled by Woofun AI shows such rapid deployment is extremely rare in blockchain history, usually seen only in networks facing existential crises. Developers said withholding the code was meant to stop attackers from reverse-engineering the flaw, not to hide technical weaknesses.   Still, this "fix first, disclose later" approach clashes with the open-source principle of public auditing, and some community members worry such emergency measures could become institutionalized and erode the network's decentralized foundation. 

Critics and Validators Clash Over the XRP Vulnerabilities Response 

On Thursday, Cyber Capital founder Justin Bons called the promotion of Ripple's decentralization to retail investors "pure fraud." He argued that the September 25 patch forced operators to upgrade and, by arriving only as a binary, undermined the community's ability to exercise oversight.  The XRPL Foundation countered that the decision followed a joint assessment of the bug's severity by the foundation, RippleX, core developers and validators, with no single party acting unilaterally. It added that version 3.4.1 was ultimately enforced through a regular two-week amendment vote backed by more than 80% of validators.   XRP Kuwait, an independent validator operator, also rejected claims that coordinated upgrades equal centralized control, saying independent operators can reach consensus through communication while keeping their autonomy.  The divide reflects two readings of decentralization: one prioritizing procedural fairness and transparency, the other outcome-driven decisions and collective security. XRPL's model proved flexible in a crisis, but it also showed how extreme pressure could pull it away from its decentralized ideals. 

The Technical Root of the XRP Vulnerabilities 

The flaw was an integer overflow bug that had existed in the XRPL payment engine since 2015. Balances are stored as 64-bit integers with a fixed maximum, and when an amount exceeds that limit, the system wraps it to a smaller value instead of returning an error.   An attacker could craft hundreds of specially priced payment requests, each seeking an enormous amount of XRP, then submit a single payment that drained them all. The requesters would be fully refunded while the payer was charged only a tiny converted sum, with the difference becoming newly minted, spendable XRP.  Exploiting these XRP vulnerabilities required carefully crafted requests and could not be triggered by ordinary transactions. Developers say no public network has shown signs of exploitation so far. This marks the first time since 2015 that XRPL has faced such a severe supply-inflation threat. Although the remediation drew criticism, it effectively reduced systemic risk.  

Splunk has patched a large set of Splunk vulnerabilities in Splunk Enterprise, the most severe of which carries a CVSSv3.1 score of 9.8. The fixes are described in two advisories, SVD-2026-1001 and SVD-2026-1002, both published on October 7, 2026, and together they cover 22 CVE identifiers, CVE-2026-76264 through CVE-2026-76285.   A security bulletin issued on October 9, 2026, rated the overall risk as medium and warned that a remote attacker could exploit some of the flaws to achieve remote code execution, denial of service, elevation of privilege, security restriction bypass, sensitive information disclosure, and data manipulation. 

Splunk Enterprise Versions Affected 

Branch	Affected versions	Fixed version
10.4	10.4.0 to 10.4.2	10.4.3
10.2	10.2.0 to 10.2.6	10.2.7
10.0	10.0.0 to 10.0.9	10.0.10
9.4	9.4.0 to 9.4.14	9.4.15

The Critical Patroni Flaw 

The top-rated issue, CVE-2026-76268 (CWE-306), stems from missing authentication in the Patroni REST API. On a search head cluster member running a version below 10.4.3 or 10.2.7, an unauthenticated attacker with network access to that interface could run operating-system commands, because critical configuration operations require no login. Versions 10.0.x and 9.4.x are not affected.   Organizations that do not use Edge Processor, OpAmp or SPL2 data pipelines can turn off the PostgreSQL sidecar by setting disabled = true in the [postgres] stanza of $SPLUNK_HOME/etc/system/local/server.conf, then restarting. 

Privilege Escalation and Secure Gateway Splunk Vulnerabilities 

CVE-2026-76266 (7.7, High) lets a local user who can run commands as the Splunk account plant content that a later Linux package upgrade executes as root. Upgrading with a tar file avoids the issue.  Three flaws hit Splunk Secure Gateway: CVE-2026-76265 (6.5) and CVE-2026-76272 (4.3) let non-admin users make it sign attacker-controlled payloads, while CVE-2026-76280 (6.3) allows writes to alert and mobile-device recipient data. They are fixed in Secure Gateway 3.10.11, 3.9.25 and 3.8.72. Admins who do not use Splunk Mobile, Spacebridge or Mission Control can disable the app instead. 

Data Exposure and Input Validation in Splunk Enterprise 

Several medium-severity bugs expose other users' data. CVE-2026-76269 (6.5) and CVE-2026-76275 (4.3) reveal other users' search jobs, CVE-2026-76270 (6.5) is a SQL injection in the SPL2 module catalog affecting only 10.4.x, and CVE-2026-76278 (4.3) leaks SPL2 module permission grants. CVE-2026-76274 (6.5), an SSRF, can disclose the Observability Cloud API token.  Others include log injection (CVE-2026-76267), a regex-based denial of service (CVE-2026-76271), source code leaking through source maps (CVE-2026-76276), collect command abuse (CVE-2026-76273, CVE-2026-76279), usernames ending in a period (CVE-2026-76277, 4.1) and scripted lookup edits by non-admins (CVE-2026-76264). 

Internally Identified Hardening Fixes 

Advisory SVD-2026-1002 groups internal findings by weakness class, scoring each CVE by its worst finding: CVE-2026-76281 (CWE-284, 9.8), CVE-2026-76284 (CWE-707, 9.0), CVE-2026-76282 (CWE-664, 8.8), CVE-2026-76283 (CWE-693, 7.6) and CVE-2026-76285 (CWE-710, 4.4). 

Fixing the Splunk Vulnerabilities 

Splunk advises upgrading Splunk Enterprise to 10.4.3, 10.2.7, 10.0.10 or 9.4.15, or later. CVE-2026-76264, CVE-2026-76265, CVE-2026-76272 and CVE-2026-76280 need extra steps. For CVE-2026-76264, admins should set scripted_lookup_raw_write_enforcement = block under [lookup] in limits.conf and restart. Where upgrading is not possible, removing run_collect from roles without internal-index access mitigates CVE-2026-76279.  Credited researchers include Gabriel Nitu of Splunk, Jean-Michel Remi Boudreau, Anton (therceman), M Mahdan Argya Syarif, Saidina Hikam, Alex Hordijk and Younes Zendour. 

The U.S. Justice Department and FBI have seized two hacking tools, Microscan and FishHub, that court documents say were operated and used by Flax Typhoon actors working for a China-based company. According to the government, the tools were used to scan, and in some cases hack, U.S. and foreign critical infrastructure systems and other networks. The court-authorized seizures were announced alongside court documents unsealed in the Western District of Pennsylvania. Prosecutors allege the actors worked for Integrity Technology Group (Integrity Tech), a company based in the People's Republic of China (PRC) that holds contracts with the PRC government. The action is the latest step by U.S. authorities against China state-sponsored hackers.

How Flax Typhoon Actors Used Microscan for Vulnerability Scanning

According to court documents, Integrity Tech built and used a botnet of internet-of-things devices infected with a variant of Mirai malware. The botnet supported the company's vulnerability scanning through Microscan, a tool Integrity Tech developed to perform reconnaissance on victim networks and identify weaknesses its clients would later exploit. Targets of Microscan scanning included a U.S. power company based in South Carolina, a multinational non-governmental organization, airports in Japan and Poland, Taiwanese critical infrastructure companies in the natural gas and power sectors, and two Taiwanese universities. Integrity Tech accessed Microscan through c0cc[.]cc, one of the seized domains.

FishHub and the Flax Typhoon Spear Phishing Operation

The second tool, FishHub, allegedly helped compromise computer networks through spear phishing. After gaining initial access, FishHub downloaded additional malware to the victim network. That malware either gave Integrity Tech's clients unauthorized remote access or searched for specific files and sent them to servers controlled by Integrity Tech. Confirmed FishHub victims included about 20 Taiwanese universities. Five seized domains helped deliver the malware: 98aicai[.]com, 98aicode[.]com, outlook3650[.]com, youtubecard[.]com, and linkedinns[.]net.

What Officials Said About Flax Typhoon

Assistant Attorney General for National Security John A. Eisenberg said the United States "will not allow China or its proxies to operate against United States interests with impunity in cyberspace." He added that the National Security Division would continue working to disrupt the Flax Typhoon threats and dismantle the infrastructure sustaining them. Assistant Director Brett Leatherman of the FBI's Cyber Division said, "The PRC relies on contractor and enabling companies to expand the reach and scale of its malicious cyber activity. By exposing and disrupting these enablers, we make it harder for the PRC to target American networks and infrastructure." U.S. Attorney Troy Rivetti for the Western District of Pennsylvania described the seizures as the second disruption of Integrity Tech's operations in as many years. Special Agent in Charge Mark Remily of the FBI San Diego Field Office said the action reflects the FBI's commitment to disrupting PRC state-sponsored cyber actors who threaten critical infrastructure and national security.

Second Disruption of Flax Typhoon Infrastructure

This is the Justice Department's second public technical disruption of Integrity Tech's hacking infrastructure. In September 2024, the department announced the court-authorized disruption of Integrity Tech's Mirai botnet, which included more than 200,000 consumer devices in the United States and worldwide. Alongside the latest seizures, the FBI and other U.S. and foreign partner agencies published a cybersecurity advisory listing indicators of compromise tied to Integrity Tech intrusion activity. The advisory is intended to help network defenders identify and respond to the group's malicious activity. The FBI San Diego and Baltimore Field Offices are investigating the case in coordination with the FBI's Cyber Division. Assistant U.S. Attorney Brendan McKenna for the Western District of Pennsylvania and Trial Attorney Jacques Singer-Emery of the National Security Division's National Security Cyber Section are prosecuting the case. Assistant U.S. Attorney Thomas Sullivan for the District of Maryland and the National Police Agency of Japan provided substantial assistance.

Atlassian has disclosed CVE-2026-21589, a critical arbitrary file access vulnerability affecting Confluence Data Center and seven other self-managed products. The advisory, published on October 5, 2026, warns that every version of the affected software is exposed and urges administrators to act right away, either by upgrading or by putting temporary safeguards in place.  Beyond Confluence Data Center, the flaw reaches Bitbucket Data Center, Jira Service Management Data Center, Jira Software Data Center, Bamboo Data Center, Crowd Data Center, Crucible and Fisheye. Atlassian is tracking the issue under BSERV-20604, CONFSERVER-104488, JSDSERVER-16809, JRASERVER-79546, BAM-26567, CWD-6610, CRUC-8741 and FE-7583. 

Which Atlassian Products CVE-2026-21589 Affects 

The bug lets an attacker with no login read specific files inside the web application's root directory. There is a catch for attackers: they must already know a file's exact name and location, since the weakness does not reveal or list directory contents. Even so, Atlassian notes that some setups may store sensitive files there, which raises the stakes.  Atlassian scores CVE-2026-21589 at 9.3, placing it in the Critical band, using the vector CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:N/VA:N/SC:H/SI:H/SA:H. The company describes this as its own internal assessment and advises organizations to judge how it applies to their environments.  Cloud users are not affected. Atlassian says its Cloud products have already been patched, its investigation found no sign of exploitation there, and Cloud customers need to take no action. 

Patched Versions for Confluence and Other Atlassian Tools 

The primary fix is an upgrade. Atlassian recommends moving each installation to a fixed release, ideally the fixed LTS version or the newest available build. 

Product	Fixed versions
Bitbucket Data Center	9.4.26, 10.2.8, 10.5.1
Confluence Data Center	9.2.26, 10.2.19
Jira Service Management Data Center	5.12.40, 10.3.26, 11.3.12
Jira Software Data Center	9.12.40, 10.3.26, 11.3.12
Bamboo Data Center	10.2.24, 12.1.12
Crowd Data Center	6.3.7, 7.0.3, 7.1.7, 7.2.4
Crucible	4.9.15
Fisheye	4.9.15

Stopgap Measures Against CVE-2026-21589 

Teams that cannot upgrade yet should take instances offline from the internet where possible, including those protected by user authentication. Atlassian then offers three mitigations, all built on one regex meant to block ".." sitting directly next to /, , or ::, including URL-encoded forms:  (?is).*(?:/|\\|::|%(?:25)*(?:2f|5c)|(?::|%(?:25)*3a){2})(?:\.|%(?:25)*2e){2}(?:/|\\|::|%(?:25)*(?:2f|5c)|(?::|%(?:25)*3a){2}|;|%(?:25)*3b|$).* 

1. Web Application Firewall rule (all products): block any URL matching the pattern in a WAF or proxy such as AWS WAF, Cloudflare or a reverse proxy, then test that encoded variants are caught. 
2. Tomcat RewriteValve (Confluence, JSM, Jira, Bamboo, Crowd): after a backup, shut down each node, enable the valve in server.xml (Crowd may use crowd.xml), and add or append a rewrite.config file in the product's WEB-INF directory. It checks both the normalized request path and the raw URI, returning a forbidden response. Restart the node afterward. 
3. urlrewrite.xml rule (Bitbucket only): back up, then place the rule at the top of /app/WEB-INF/urlrewrite.xml on every node, mirror and mirror farm node, returning a 404 via /mvc/error404. Restart Bitbucket. 

Checking Confluence Logs for Signs of Compromise 

Atlassian says it cannot confirm whether any customer instance has been hit and advises bringing in local security teams. Investigators can URL-decode each access-log request up to twice and search for ".." next to /, , or ::, or run the regex directly against raw log lines.  Under Atlassian's security bug fix policy, critical fixes are backported as new maintenance releases instead of binary patches. Customers can subscribe to alert emails at my.atlassian.com/email and raise questions through support.atlassian.com. 

The FBI and the U.S. Secret Service (USSS) have issued a joint cybersecurity advisory warning organizations about FortiBleed, an active global credential-compromise campaign targeting internet-facing Fortinet FortiGate firewalls and SSL VPN gateways. The advisory, published on October 6, 2026, cites SOCRadar data verifying more than 86,644 compromised devices across 194 countries.

According to the advisory, the campaign exploits reused or leaked credentials and legacy SHA-256 password storage, allowing threat actors to harvest and crack authentication data at scale. Attackers continue to scan exposed Fortinet firewalls using previously obtained credentials, and some affected organizations have been locked out of their own systems.

How the FortiBleed Attack Chain Works

The operation's internal workflow came to light after the threat actors unintentionally exposed their own backend server. The open directory revealed a mature, multi-stage campaign run as an initial access broker operation.

Attackers first scanned the internet for exposed FortiGate SSL VPN portals using automated scripts. They then gathered credentials through credential stuffing and password spraying, drawing on prior Fortinet leak dumps and infostealer logs.

Password hashes taken from compromised devices were exfiltrated and fed into a GPU-accelerated cracking cluster running on rented infrastructure, where Hashcat and Hashtopolis ran distributed jobs to convert stolen data into plaintext credentials.

[caption id="attachment_114628" align="aligncenter" width="800"] Source: FBI[/caption]

Cracked credentials were then validated, with scripts filtering out honeypots, mapping organizations and ranking high-value targets by revenue and network structure. Attackers created new administrative accounts on firewalls to maintain persistence, then moved into victim networks to enumerate Active Directory and identify privileged accounts. The final stage involved selling access, including working VPN configurations and target lists, to downstream threat actors.

FortiBleed Lockouts and Ransomware Links

The advisory warns that victims may lose access to their Fortinet devices if attackers delete original accounts or change their passwords. In certain cases, threat actors removed existing accounts to block organizations from their devices while attempting lateral movement. Recovery may require steps beyond standard patching and password resets.

Reporting also indicates that initial access brokers using the FortiBleed attack chain have supplied access to ransomware affiliates, currently including INC/Lynx ransomware and Payload ransomware.

FortiBleed Indicators of Compromise

The agencies released indicators of compromise (IOCs) to help defenders review their environments. IP address 45.154.12.132 was observed as a command and control server, while 154.202.59.169 and 103.27.186.156 served as proxy nodes and 45.155.250.158 functioned as a beacon relay. The address 85.11.187.8 was linked to the Hashtopolis tool. The advisory also lists 13 IP addresses observed between June 18 and July 23, 2026, conducting brute force attacks or authenticating with compromised accounts.

Account names found on victim systems include adminin, fortiAdmin, forticloud-sync, fgtsecure, adminsslvpn, itadmin and roadmin. The agencies note that IP addresses may be reassigned over time and advise organizations to vet IOCs before acting on them.

Recommended Mitigations Against FortiBleed

The FBI and USSS recommend that organizations:

• Restrict external management through trusted hosts, a local-in policy, or by removing internet administration entirely.
• Terminate all active admin and VPN sessions and reset all Fortinet VPN and administrative passwords.
• Enforce phishing-resistant MFA on all remote access and administrative accounts.
• Review firewall and VPN configurations for unrecognized accounts and unauthorized changes.
• Check firewall, VPN, authentication and domain controller logs for unusual access.
• Use PBKDF2 to store administrator credentials and remove weaker legacy hashes, following Fortinet guidance for FortiOS v7.2.11 and later.
• Review all REST API keys, remove unknown ones and refresh legitimate keys.

Organizations that detect a compromise are urged to isolate affected hosts, scope the intrusion and report incidents to the FBI's Internet Crime Complaint Center (IC3), a local FBI field office or a local USSS field office.

Google has suspended its Open Source Software Vulnerability Rewards Program (OSS VRP), a vulnerability search initiative that paid researchers for finding flaws in the company's open-source software. The pause took effect on October 1, 2026. Google blamed a "significant rise" in automated and AI-generated reports, most of which turned out to be invalid.  Google announced the suspension in posts on X and on the program's website. It said it would provide "an update" in the first quarter of 2027. According to Tom's Hardware, Google engineers and open-source maintainers had been overwhelmed by reports that were invalid or contained hallucinations. 

Why Google Halted OSS VRP 

"This pause is due to a significant rise in automated submissions, the vast majority of which are not valid," the company said. In the meantime, Google is encouraging participants to look at its other bug bounty programs.  As of October 1, 2026, the OSS VRP no longer accepts product vulnerability reports. Some reports involving Google Cloud repositories that affect Cloud products may still be accepted through the Cloud VRP. Flaws in open-source projects closely tied to Cloud or AI products can go to the Cloud VRP or the AI VRP. Researchers can also turn to the Patch Rewards Program, which pays for security improvements to Google's open-source projects. Reports submitted before October 1 are not affected.  Product vulnerabilities previously covered flaws that substantially affected the confidentiality or integrity of user data. Examples included memory corruption in file format parsers or network protocols, sanitizer failures, path traversal, and insecure defaults or code examples in documentation.   For OT0 and OT1 repositories, memory corruption reports required exact OSS-Fuzz reproduction steps or an already merged patch. Non-memory corruption reports did not need a patch. Google also urged researchers to build OSS-Fuzz integrations for important projects that lacked them before reporting. 

How Project Tiers Shape Vulnerability Search Rewards 

Repositories fall into four tiers: 

• Flagship (OT0) projects are the most sensitive and carry the highest payouts. 
• Important (OT1) projects have significant community impact or security criticality. Both OT0 and OT1 are listed at github.com/google/bughunters. 
• Standard (OT2) covers active repositories with stable packages on registries such as npm or PyPI. There is no published list for this tier. 
• Low-priority (OT3) includes experimental, sample, research, and archived projects. These earn no financial rewards. 

Supply chain compromises pay $3,133.7 to $31,337 for OT0, $1,337 to $13,337 for OT1, and $500 to $3,133.7 for OT2. Other security issues, such as leaked write-access credentials or weak passwords on third-party CI systems, earn $1,000 for OT0 and $500 for OT1. OT2 and OT3 projects receive nothing for product vulnerabilities or other security issues.  A reward panel sets final amounts based on security impact alone. It may pay more for clever or wide-reaching bugs and less for those that depend on hypothetical flaws. It typically pays once per root cause, though bonuses of about $1,000 are possible for especially clever findings or well-written reports. Researchers can donate their rewards to charity. Rewards left unclaimed after 12 months go to a charity of Google's choosing.  Several categories are excluded: 

• Issues rooted in downstream integration 
• Typosquatting without proof that build artifacts were compromised 
• Insecure installation instructions 
• Social engineering 
• Bugs with negligible impact 

Google also asks researchers to avoid DoS attacks, black hat SEO, spam, and high-traffic automated testing tools. Reports go through Google's vulnerability form and should include a buildable proof of concept, the affected versions, an impact description, and an attack scenario. 

This weekly roundup covers a customer data breach at Spain's national rail operator, an Australian government warning about attackers hijacking corporate AI services, a patient data theft from a Polish medical software platform, phishing campaigns that turn legitimate remote management tools against their victims, and a gathering of Europe's police leaders focused on how AI is reshaping crime.  Two threads run through this week's stories. Artificial intelligence is now both a target and a tool for attackers: stolen API keys are draining company budgets, and AI is reportedly being used in intrusions. Meanwhile, attackers keep exploiting trusted software, from a medical records platform to legitimate IT administration tools, to stay hidden for days or weeks. 

The Cyber Express Weekly Roundup 

Renfe Confirms Data Breach Through Compromised Adif Servers 

Spanish rail operator Renfe confirmed on September 25 that hackers had accessed customer names and email addresses. The intrusion came through compromised servers belonging to Adif, the state-owned manager of Spain's rail infrastructure. Renfe said no financial data was exposed and train services were not disrupted. Read more… 

Australia Warns of Stolen AI Keys as Hijacked Access Costs Firms Up to $600,000 

The Australian Signals Directorate (ASD) has warned that attackers are taking over organizations' AI services. They get in through exposed API keys, stolen sessions, and compromised suppliers. Once inside, they burn through paid credits and generate harmful content. Where AI agents are deployed, they also use that access to reach connected systems and data. Read more… 

Polish Patient Data Stolen in Cyberattack on Medyc Medical Software 

Attackers exploited an SQL injection vulnerability in Medyc, a medical records platform developed by Polish vendor Qbusoft, on August 22–23. They stole patient names, PESEL national identification numbers, addresses, phone numbers, and email addresses going back to July 2024. The breach went unnoticed until September 8–9, more than two weeks later. Read more… 

Phishing Campaigns Abuse MSP360 Remote Management Tool to Install ScreenConnect 

Microsoft has tracked phishing campaigns that tricked victims across several industries into installing MSP360, a legitimate remote management tool. The installers were disguised as Zoom downloads, Adobe updates, invitations, and delivery notices. After gaining a foothold, the attackers installed ConnectWise ScreenConnect as a backup access channel. Read more… 

Europe's Police Chiefs Warn AI Is Making Crime Faster and Larger in Scale 

More than 500 police leaders from 59 countries gathered at Europol's headquarters from September 28 to 30 to discuss how AI is changing crime. Europol warned that criminals are using generative AI, deepfakes, and autonomous agents to expand fraud, cybercrime, migrant smuggling, and child exploitation. It also pointed to AI's potential to speed up police investigations. Read more… 

Weekly Cybersecurity Takeaway 

Artificial intelligence features in several of this week's stories, in different roles. In Australia, AI services are the target: attackers hijack them to steal paid credits and to reach connected systems through AI agents. At Europol's gathering, police leaders described AI as a force multiplier that lets criminal groups run fraud and exploitation at greater speed and scale. In the Renfe case, AI was reportedly used in the attack, though that remains unconfirmed.  The other common thread is how attackers abuse what organizations already trust. In Poland, an injection flaw in widely used medical software exposed sensitive patient identifiers and went undetected for more than two weeks. Microsoft's findings show attackers installing two legitimate remote management tools so their activity looks like ordinary IT work.   The Renfe breach, which reached customers through a partner organization's servers, is a reminder that an organization's exposure extends to its suppliers and infrastructure partners.  Together, these developments call for practical steps. Organizations should treat AI credentials as carefully as any other privileged secret. They should watch for unexpected remote access software even when it is legitimate, and they should hold the vendors who handle their data to the same security standards they apply internally. 

Google is limiting access to Gemini 4 Argon, its new artificial intelligence model capable of autonomously finding, validating and patching critical software vulnerabilities, to a vetted group of cybersecurity experts. The restricted rollout is intended to give defenders access to the model's capabilities while Google strengthens safeguards against potential misuse by malicious actors. Koray Kavukcuoglu, SVP of Google DeepMind and Chief AI Architect at Google, announced the model on September 30, 2026. Google is initially distributing it through its Fairwind Program and plans to expand access after gathering feedback from trusted testers and refining its security measures. The decision puts the focus on a central cybersecurity challenge: AI that can help security teams identify and fix weaknesses could also provide capabilities that attackers might misuse to discover and exploit vulnerabilities.

Gemini 4 Argon Discovers Critical Software Vulnerabilities

Google says Gemini 4 Argon can identify, validate and patch vulnerabilities across complex software environments. Its reported capabilities include examining codebases, identifying potential exposures and producing evidence to validate security weaknesses. In an early deployment, cybersecurity company Wiz used Argon through its Scan for Good initiative, which focuses on identifying and remediating high-risk exposures affecting critical public infrastructure. According to Google, the model discovered a critical vulnerability that exposed sensitive personal information in healthcare software used by hospitals worldwide. The company said previous frontier AI models had missed the issue but did not publicly identify the affected software or provide technical details about the flaw. On CWE-bench v1, a benchmark designed to evaluate vulnerability remediation, Argon tied for first place with a score of 68%. Google also reported improvements over its earlier 3.8 Flash Cyber model in internal testing. On Wiz's internal black-box penetration testing benchmark, Argon performed better at identifying attack surfaces, discovering vulnerabilities and generating proof-of-concept evidence to validate its findings. These results are based on Google's and its partners' reported evaluations. They do not establish how reliably the model will identify vulnerabilities across every real-world environment.

Why Google Is Limiting Access to Gemini 4 Argon

Google is taking a phased approach because the model's advanced capabilities require additional safeguards before wider deployment. For defenders, automated vulnerability discovery could help identify weaknesses earlier and reduce the time needed to investigate and remediate them. However, similar capabilities could also be misused to locate weaknesses in systems that attackers want to compromise. Google said trusted cyber defenders and its own internal teams will receive access to Argon without cyber-specific guardrails so they can use its full defensive capabilities. The company is also participating in the US government's voluntary process for pre-release AI model access. Google has not announced a specific date for general availability. It said it will continue evaluating feedback from early testers while strengthening protections against misuse.

Prompt Injection Attacks Among Key Security Concerns

Google has identified several areas where it is strengthening safeguards before expanding access to Gemini 4 Argon. One is prompt injection attacks, in which malicious instructions hidden in external content attempt to manipulate an AI model's behavior. Google said Argon demonstrated improved resilience against indirect prompt injection on Gray Swan's Indirect Prompt Injection benchmark. The company is also testing protections against requests that could facilitate cyberattacks or the development of chemical, biological, radiological and nuclear threats. Google said internal and external red teams assessed these safeguards using manual and automated attack methods. Additional measures monitor the model's reasoning and actions for potential misalignment, with mechanisms designed to stop execution when necessary. Google is also hardening isolated testing environments used for high-risk training and evaluations. These measures are intended to reduce risks, but the announcement does not establish that every possible misuse scenario or security failure has been eliminated.

Gemini 4 Argon Availability and Pricing

Google plans to expand access to Gemini 4 Argon after gathering feedback and continuing its safeguards work. Paid API customers and Google AI Ultra subscribers are expected to be among the first groups to receive broader access. The announced introductory API pricing is $2 per million input tokens and $10 per million output tokens. Cached input tokens receive a 95% discount on the standard input price. For cybersecurity teams, the key issue is whether the model's reported vulnerability discovery capabilities translate into reliable results across production systems. Its wider rollout will also test whether safeguards can keep pace with AI systems capable of supporting both vulnerability remediation and offensive security techniques.

OpenSSL has disclosed a high-severity vulnerability in its Datagram Transport Layer Security (DTLS) implementation. The flaw could let a remote peer read unintended plaintext heap memory during handshake data transmission, or cause a denial-of-service condition. Tracked as CVE-2026-84782, it stems from an out-of-bounds read in how DTLS handshake message retransmissions are handled.  According to the Security Advisory published on September 29, 2026, the bug affects OpenSSL versions 4.0, 3.6, 3.5, 3.4, 3.0, 1.1.1, and 1.0.2. 

How CVE-2026-84782 Breaks OpenSSL's DTLS Retransmission Logic 

DTLS is built for datagram-based transports, where packets can go missing, so it includes retransmission mechanisms to recover lost handshake packets. That recovery feature is where the problem lies.  The trouble starts when an application sends a DTLS handshake message in several fragments and the transport temporarily stops accepting data. OpenSSL may then return a WANT_WRITE status, pausing the handshake write partway through. While the write is suspended, a DTLS retransmission timer can expire and try to resend an earlier handshake message from the retransmission queue.  At that point, OpenSSL's retransmission code wrongly reuses the internal buffer and tracking state tied to the paused write. Instead of resetting its read position to the start of the queued message, it may begin at the offset the in-progress write had already reached.  As a result, a peer can receive a malformed or mislabeled DTLS handshake message containing leftover bytes from an unrelated, larger message still in transit. In some cases, the read extends past the intended buffer, exposing heap-resident data in plaintext. 

Memory Disclosure and Crash Risks 

The weakness is classified as CWE-125, an out-of-bounds read, and its main security impact is unintended disclosure of heap memory to the remote DTLS peer. What leaks depends on the process's memory contents and runtime conditions, and it could include fragments of previously allocated application or library data. The advisory, however, has not confirmed any reliable arbitrary-memory disclosure capability.  The bug can also crash a process if the retransmission logic reads from an unmapped memory region, creating a denial-of-service risk for applications that handle attacker-controlled DTLS traffic.  A related issue in the same retransmission path can corrupt the bookkeeping needed to resume the suspended handshake write. If a retransmission occurs while another write is still paused, later calls to SSL_read(), SSL_write(), SSL_accept(), or SSL_connect() may run into inconsistent states, triggering an abort in debugging builds. 

OpenSSL Releases Fixes for the DTLS Vulnerability 

OpenSSL fixed the problem by resetting the retransmission read position before a message is resent. The patch also skips retransmissions while a handshake write is suspended, deferring them until the application resumes the pending operation. 

Affected branch	Fixed release
OpenSSL 4.0	4.0.3
OpenSSL 3.6	3.6.5
OpenSSL 3.5	3.5.9
OpenSSL 3.4	3.4.8
OpenSSL 3.0 Premium Support	3.0.23
OpenSSL 1.1.1 Premium Support	1.1.1zj
OpenSSL 1.0.2 Premium Support	1.0.2zs

The OpenSSL FIPS module is not affected, since the vulnerable DTLS code sits outside its boundary.  Organizations are advised to identify internet-facing services, VPNs, VoIP systems, IoT deployments, and custom applications that rely on DTLS, and to promptly upgrade their bundled or system-provided OpenSSL packages.  Laurent Gaffie of SecuRizon reported CVE-2026-84782 on August 17, 2026, and Ryan Hooper developed the fix. 

Apple has released iOS 26.7.1 and iPadOS 26.7.1 to address CVE-2026-86950, a CoreGraphics vulnerability the company says may have been used in targeted attacks. The updates came out on September 28, 2026, along with matching fixes for macOS Tahoe 26.7.1 and macOS Sequoia 15.8.1.  The flaw is in CoreGraphics, a core Apple framework that handles graphics rendering. According to Apple, processing a maliciously crafted file may lead to arbitrary code execution. In practice, an attacker could run their own code on a vulnerable device by getting it to open a specially prepared file.  Apple described the root cause as an out-of-bounds write issue. This type of memory bug happens when software writes data past the edge of the memory area set aside for it. The company said it fixed the problem with improved bounds checking.  In its advisory, Apple stated: "Apple is aware of a report that this issue may have been exploited in an extremely sophisticated attack against specific targeted individuals on versions of iOS before iOS 27."  Apple credited Meta Product Security with reporting the vulnerability. 

Devices Covered by the iOS and iPadOS Update 

The iOS 26.7.1 and iPadOS 26.7.1 fix is available for iPhone 11 and later. On the iPad side, it covers: 

• iPad Pro 12.9-inch (3rd generation and later) 
• iPad Pro 11-inch (1st generation and later) 
• iPad Air (3rd generation and later) 
• iPad (8th generation and later) 
• iPad mini (5th generation and later) 

CVE-2026-86950 Also Reaches macOS 

The issue is not limited to iOS and iPadOS. Apple published separate security notes for macOS Tahoe 26.7.1 and macOS Sequoia 15.8.1, both released on September 28, 2026. Each describes the same CoreGraphics flaw, the same arbitrary code execution risk, and the same out-of-bounds write fix. Each also credits Meta Product Security.  The macOS notes repeat Apple's statement that the bug may have been exploited against specific individuals on versions of iOS before iOS 27. The Tahoe update applies to macOS Tahoe, and the Sequoia update applies to macOS Sequoia.

Advisory Rates CVE-2026-86950 as High Risk 

A separate advisory, also dated September 28, 2026, classifies the problem as a remote code execution vulnerability in Apple products. It warns that a remote attacker could exploit the flaw to trigger remote code execution on a targeted system.  The advisory rates the risk level as High for two reasons. The flaw is being exploited in an extremely sophisticated attack on specific targeted individuals running versions of iOS before iOS 27. A crafted file is also enough to achieve arbitrary code execution.  The advisory lists the following as affected: 

• Versions prior to iOS 26.7.1 and iPadOS 26.7.1 
• Versions prior to macOS Tahoe 26.7.1 
• Versions prior to macOS Sequoia 15.8.1 

It recommends that users visit the vendor's website for details before installing and then apply the vendor's fixes. The related Apple support pages are: 

• https://support.apple.com/en-hk/149226 
• https://support.apple.com/en-hk/149228 
• https://support.apple.com/en-hk/149229 

Apple's Disclosure Approach 

Apple says that, to protect customers, it does not disclose, discuss, or confirm security issues until it has investigated them and patches or releases are available. The company references vulnerabilities by CVE-ID when possible, as it did with CVE-2026-86950. Recent releases are listed on its security releases page, and more information is on its Product Security page.  Apple's notes also state that information about products not made by Apple, or about independent websites it does not control, is provided without recommendation or endorsement. Apple directs users to contact those vendors for further details.  Since attackers have reportedly already used the flaw, owners of supported iPhones and iPads should install iOS 26.7.1 or iPadOS 26.7.1. Mac users should update to macOS Tahoe 26.7.1 or macOS Sequoia 15.8.1, depending on which version they run. 

This weekly roundup covers a brazen breach claim against the FBI's recruitment portal, a multinational advisory exposing North Korea's fake-recruiter malware operation, a sweeping EU proposal to reshape children's access to social media, a conversation on application security in the age of AI agents, a short-lived Discord ban in the Philippines, and a multimillion-dollar hot-wallet heist at a crypto casino.  From federal law enforcement to individual job seekers, this week's developments show that no target is too prominent or too small. Cybercriminals are openly taunting the agencies that pursue them, state-backed actors are turning the hiring process into an attack vector, and governments are testing how far they can push platforms on child safety, sometimes reversing course within hours. 

The Cyber Express Weekly Roundup 

Shiny Hunters Claims FBI Breach, Says FBIjobs Agent Data Exposed 

The FBI is investigating an apparent breach after the cybercriminal group Shiny Hunters claimed to have stolen personal data belonging to thousands of federal agents, with the probe centering on the bureau's FBIjobs recruitment website. Read more... 

WaterPlum Hackers Steal $10.7M in Crypto From IT Workers 

A North Korean cyber actor group known as WaterPlum, also referred to as "Contagious Interview," has infected at least 30,000 devices across more than 100 countries while stealing cryptocurrency from IT professionals worldwide. Read more... 

EU KIDS Act Sets New Social Media Rules for Children 

The European Commission adopted the proposal on September 18, setting out rules that would prevent children under 13 from accessing social media and establish 15 as the minimum age for minors to open accounts independently. Users aged 13 to under 15 could instead access parent-managed mini accounts with limits on social contacts and up to 60 minutes of screen time per day, while platforms would bear the responsibility of proving their services are safe for children. Read more… 

Harness' Rahul Sood: AppSec in the Agentic Era Is About More Than Code. It's About Authority 

In this conversation, The Cyber Express spoke with Rahul Sood, the General Manager for Application Security at Harness, who leads the company's AppSec portfolio and focuses on building security for the AI era, integrated directly into modern DevOps workflows. The conversation centers on a shift in how security teams must think about AI agents: the risk no longer lies only in the code being shipped, but in what authority autonomous agents are granted once they operate inside production systems. Read more… 

Ban on Discord Lifted After Platform Commits to Work With DICT and CICC 

The Philippine government has lifted its ban on Discord after the platform's representatives met with officials from the Department of Information and Communications Technology (DICT) and the Cybercrime Investigation and Coordinating Center (CICC), with the restriction removed on Thursday, Sept. 24. Read more... 

Duelbits Hit by $7 Million Hack as Crypto Stolen Across Four Blockchains 

Crypto casino Duelbits has shut down its platform after attackers took roughly $7 million from several of its wallets, with the stolen assets moved across four blockchains and mostly converted into Ether. The company says funds held in user accounts have not been affected, but it has not explained how the attackers got in or when the site will return. Read more... 

Weekly Cybersecurity Takeaway 

This week's developments span the full spectrum of cyber risk, from a criminal group openly retaliating against the FBI over a public advisory, to a state-backed operation weaponizing job interviews against developers, to a crypto platform losing millions from the very hot wallets that keep it running.  Meanwhile, the policy response is moving on multiple fronts: the EU is flipping the burden of proof onto platforms to demonstrate child safety by design, and the Philippines has shown how quickly a government can block, and then restore, a major platform in exchange for enforcement commitments. At the same time, security leaders are warning that AI agents are introducing an entirely new question of who, or what, holds authority inside enterprise systems.  Taken together, the stories point to an increasingly confrontational threat landscape, where attackers are bolder, trust-based processes like hiring are being exploited at scale, and institutions are experimenting with more aggressive levers to hold platforms and adversaries accountable. 

Artificial intelligence systems are becoming more capable and autonomous, raising new questions about how they can be monitored, controlled and secured. Speaking before the United Nations Security Council, OpenAI CEO Sam Altman warned that the rapid pace of AI development could create significant AI security risks, particularly if systems advance faster than institutions can respond. Altman highlighted concerns ranging from human control and autonomous AI to emerging vulnerabilities, concentrated technological power and the need for international incident reporting. He also called for stronger cooperation between governments, critical infrastructure operators and technical experts to address emerging AI threats.

Here are six key AI security risks highlighted in his remarks.

1. Losing Human Control Over Advanced AI

One of Altman’s central concerns was the possibility that humans could lose control over increasingly capable AI systems. He warned that AI could move so quickly that people may no longer be able to understand what systems are doing or intervene when necessary. He argued that developers should not train systems unless they can make a strong case that those systems can remain under human control. The concern becomes more significant as AI systems become capable of handling increasingly complex tasks and making decisions with less direct human involvement.

2. AI Systems Becoming More Autonomous

The growing autonomy of AI systems presents another challenge for AI cybersecurity. Altman pointed to the possibility of AI systems improving themselves and future versions of themselves, describing this as recursive self-improvement. As the process of developing AI becomes more automated, the pace of progress could accelerate. For security teams and policymakers, increasingly autonomous systems could make monitoring, testing and intervention more difficult.

3. Emerging Vulnerabilities and New Threats

More capable AI systems could also introduce vulnerabilities and threats that organizations and governments may not yet fully understand. Altman called for secure channels between governments, critical infrastructure operators and technical experts to share information about emerging vulnerabilities and threats. The remarks place information sharing at the center of preparing for AI vulnerabilities, particularly as AI becomes increasingly integrated into important systems and infrastructure.

4. AI Power Concentrating in Too Few Hands

Altman also warned about the concentration of advanced AI capabilities among a small number of companies, individuals or countries. He argued that no single company or country should be able to use the most powerful AI systems to impose its worldview on others. From a security perspective, concentrated control over highly capable AI could create systemic risks if a limited number of actors have disproportionate access to powerful technology.

5. The Need for Global AI Incident Reporting and Threat Sharing

Altman called for accurate and speedy AI incident reporting, including classification and reporting protocols that would allow countries to learn from failures before they become larger problems. He also called for common standards covering AI capabilities, risk assessments, safeguards and human oversight. Such standards, he said, could help countries compare evidence and establish a shared understanding of what effective safeguards and oversight should look like.

6. AI Development Could Outpace Existing Safety Measures

The speed of AI development itself could become a security challenge. Altman warned that rapid progress could move faster than institutions are prepared to handle. He argued that competition between companies should not become a reason to accept excessive technological risk. He also stressed the importance of alignment, monitoring and safety guarantees as AI systems become more capable. For AI safety, this means ensuring that security measures and oversight develop alongside increasingly powerful systems rather than being added only after problems emerge.

AI Security Moves Into a Global Conversation

Altman’s remarks put AI security risks alongside broader questions about international cooperation, human oversight and technological governance. His proposed approach includes global standards for measuring AI capabilities and risks, incident reporting mechanisms, and secure communication between governments, critical infrastructure operators and technical experts. As AI systems become more autonomous and capable, the challenge will not only be what these systems can do, but whether organizations and governments can monitor, understand and respond to the risks they create.