Why Should I Care? β€” 2026-11-14 | πŸ”΄ 2 HIGH Β· 🟑 14 MEDIUM Β· πŸ”΅ 15 RADAR Β· βšͺ 36 FILTERED

πŸ“‹ Briefing β€” 2026-11-14

67 vendor intel items scanned  |  πŸ”΄ 2 HIGH  |  🟑 14 MEDIUM  |  πŸ”΅ 15 RADAR  |  βšͺ 36 FILTERED

πŸ”΄ Critical β€” action required:

  1. Self-service Security with vDefend and VCF Automation β€” Skip this if you are looking for a vulnerability fix; this is a feature announcement for VMware Cloud Foundation 9.1 that introduces self-service security delegation for tenant administrators.
  2. CVE-2026-119988 Windows SMB Remote Code Execution Vulnerability (CVE-2026-119988) β€” Skip this if you are not running Windows; this is a Windows SMB vulnerability, but the provided advisory text is an informational update with no technical details.

Everything else can wait.

🟑 Medium β€” review when time permits:

  1. Amazon Threat Intelligence observes threat actors running cybercrime campaigns with ARTEX AI agents β€” Yes, if you run any of the listed AI gateway, proxy, or application builder platforms (e.g., LiteLLM, One API, Dify, Langflow, RAGFlow) exposed to the internet: you are a primary target for automated credential theft and cloud privilege escalation.
  2. Chromium CVE-2026-125342: Uninitialized resource (+ 35 weitere Chromium CVEs im gleichen Build) β€” Yes, if you run Microsoft Edge or any Chromium-based browser: this is an uninitialized resource vulnerability that can lead to crashes or potential code execution, and it is part of a batch of 36 fixes in the latest stable release.
  3. CVE-2026-113749 Microsoft Windows Codecs Library Remote Code Execution Vulnerability β€” Yes, if you run any supported version of Microsoft Windows or Microsoft Office: this is a Remote Code Execution (RCE) vulnerability in the Windows Codecs Library that allows attackers to execute arbitrary code on your system.
  4. CVE-2026-95948 Windows Graphics Kernel Elevation of Privilege Vulnerability β€” Yes, if you run Windows: This is an Elevation of Privilege (EoP) vulnerability in the Windows Graphics Kernel that allows a local attacker to gain SYSTEM-level control of your machine.
  5. CVE-2026-95932 Windows Graphics Kernel Elevation of Privilege Vulnerability β€” Yes, if you run Windows 10, 11, or Server 2016/2019/2022: This is a local privilege escalation vulnerability in the Windows Graphics Kernel that allows an attacker with local access to gain SYSTEM privileges.
  6. CVE-2026-88217 Windows Graphics Kernel Information Disclosure Vulnerability β€” Vendor advisory for an information disclosure vulnerability in Windows Graphics Kernel with no active exploitation reported.
  7. CVE-2026-96150 Microsoft Windows Codecs Library Information Disclosure Vulnerability β€” Vendor advisory for an information disclosure vulnerability in Windows Codecs Library with no active exploitation reported.
  8. CVE-2026-96144 Windows Graphics Kernel Information Disclosure Vulnerability β€” Vendor advisory for an information disclosure vulnerability in Windows Graphics Kernel with no active exploitation reported.
  9. CVE-2026-96126 Windows Graphics Kernel Elevation of Privilege Vulnerability β€” Vendor advisory for an elevation of privilege vulnerability in Windows Graphics Kernel with no active exploitation reported.
  10. CVE-2026-95953 Windows Graphics Kernel Elevation of Privilege Vulnerability β€” Vendor advisory for an elevation of privilege vulnerability in Windows Graphics Kernel with no active exploitation reported.
  11. ... and 4 more

πŸ”΅ 15 items on the radar β€” see below ↓


πŸ”΄ HIGH β€” Handle Now


Self-service Security with vDefend and VCF Automation

VMware / Broadcom Security

❓ Why Should I Care?
Skip this if you are looking for a vulnerability fix; this is a feature announcement for VMware Cloud Foundation 9.1 that introduces self-service security delegation for tenant administrators.

🎯 Affected versions: N/A (Feature Release)
Not affected: N/A

🎭 In plain English:
This is not a security flaw but a new capability in VMware Cloud Foundation 9.1 that allows tenant administrators to manage their own firewall rules without waiting for central IT. For example, a developer team can now configure their own network isolation policies within the guardrails set by the provider, similar to how they would manage security groups in AWS or Azure.

πŸ”§ Prerequisites:

  • Running VMware Cloud Foundation 9.1
  • Utilizing VCF Automation
  • Provider administrator has delegated vDefend security roles to tenant administrators

⏱ Urgency: No immediate action is required as this is a new feature release, but organizations should review their RBAC delegation models to ensure tenants are not granted excessive security control.

βœ… Fixed in: 9.1

πŸ’‘ Context: The root cause is the traditional bottleneck of centralized security management, which this feature addresses by enabling delegated, self-service security governance within defined guardrails.


CVE-2026-119988 Windows SMB Remote Code Execution Vulnerability

Microsoft MSRC | CVE-2026-119988

❓ Why Should I Care?
Skip this if you are not running Windows; this is a Windows SMB vulnerability, but the provided advisory text is an informational update with no technical details.

🎯 Affected versions: Windows (specific versions not listed in provided text)

🎭 In plain English:
This is a remote code execution flaw in the Windows file sharing protocol (SMB). An attacker could potentially send a malicious file or request to a Windows machine to execute code without the user's knowledge, potentially taking full control of the system.

πŸ”§ Prerequisites:

  • Running a vulnerable version of Windows
  • SMB service enabled and accessible by the attacker

⏱ Urgency: Urgency is moderate because the advisory text provided is an informational update without specific exploit details, but SMB RCEs are generally high-risk.

πŸ’‘ Context: The provided text is an informational update acknowledging the vulnerability but does not specify the root cause or affected versions.


🟑 MEDIUM (14)


Amazon Threat Intelligence observes threat actors running cybercrime campaigns with ARTEX AI agents

AWS Security Blog

❓ Why Should I Care?
Yes, if you run any of the listed AI gateway, proxy, or application builder platforms (e.g., LiteLLM, One API, Dify, Langflow, RAGFlow) exposed to the internet: you are a primary target for automated credential theft and cloud privilege escalation.

🎯 Affected versions: LiteLLM (default master key), One API family (isCliRequest header bypass), Langflow (CVE-2025-3248), RAGFlow (CVE-2026-28797), Dify, FastGPT, Langflow, RAGFlow, Open WebUI, LobeChat, Ollama, New API, Veloera, Sub2API

🎭 In plain English:
Threat actors are using AI agents to automatically find and steal API keys from AI management tools that are left open to the internet. For example, an attacker's AI agent scans the web, finds a LiteLLM instance with its default password, steals the master key, and uses that key to access expensive AI models or cloud resources, effectively stealing your compute budget and data.

πŸ”§ Prerequisites:

  • Target AI gateway or application builder is exposed to the internet
  • Default credentials or uninitialized installation are present
  • Known vulnerabilities (e.g., CVE-2025-3248, CVE-2026-28797) or authentication bypasses (e.g., isCliRequest header) are unpatched
  • Secrets are stored in client-side JavaScript, source maps, or environment files accessible to the agent

⏱ Urgency: Immediate action is required because these campaigns are automated, self-sustaining, and actively harvesting credentials to fund further attacks.

βœ… Fixed in: Latest stable release of LiteLLM, Latest stable release of One API, Latest stable release of Langflow (patching CVE-2025-3248), Latest stable release of RAGFlow (patching CVE-2026-28797), Latest stable release of Dify, Latest stable release of FastGPT, Latest stable release of Open WebUI, Latest stable release of LobeChat, Latest stable release of Ollama

πŸ’‘ Context: The root cause is the combination of internet-exposed AI management interfaces with weak default security configurations and known, unpatched vulnerabilities that allow automated agents to extract upstream provider keys.


Chromium CVE-2026-125342: Uninitialized resource (+ 35 weitere Chromium CVEs im gleichen Build)

Microsoft MSRC | CVE-2026-125342

❓ Why Should I Care?
Yes, if you run Microsoft Edge or any Chromium-based browser: this is an uninitialized resource vulnerability that can lead to crashes or potential code execution, and it is part of a batch of 36 fixes in the latest stable release.

🎯 Affected versions: Microsoft Edge versions prior to the latest stable release (specific version numbers not listed in advisory, but implied to be all versions ingesting the vulnerable Chromium build)
Not affected: Microsoft Edge versions updated to the latest stable release containing the fix

🎭 In plain English:
The browser has a bug where it tries to use a piece of memory that hasn't been properly set up yet. An attacker could exploit this by tricking the browser into processing a malicious web page, which could cause the browser to crash or potentially run malicious code on your computer.

πŸ”§ Prerequisites:

  • Running a vulnerable version of Microsoft Edge or Chromium-based browser
  • Attacker delivers malicious content (e.g., via a website or email link) to the user
  • User interacts with the malicious content

⏱ Urgency: High urgency because this is part of a large batch of critical fixes in the latest stable release, and unpatched browsers are exposed to active exploitation risks.

βœ… Fixed in: Latest stable Microsoft Edge release

πŸ’‘ Context: The root cause is an uninitialized resource in the Chromium engine, which was fixed in the latest stable Chromium build ingested by Microsoft Edge.


CVE-2026-113749 Microsoft Windows Codecs Library Remote Code Execution Vulnerability

Microsoft MSRC | CVE-2026-113749

❓ Why Should I Care?
Yes, if you run any supported version of Microsoft Windows or Microsoft Office: this is a Remote Code Execution (RCE) vulnerability in the Windows Codecs Library that allows attackers to execute arbitrary code on your system.

🎯 Affected versions: All supported versions of Microsoft Windows and Microsoft Office (specific versions not listed in provided text, but implied to be all supported releases prior to the patch)

🎭 In plain English:
The part of Windows that handles image and media files has a flaw that lets a malicious file trick your computer into running harmful instructions. An attacker could send you a specially crafted image or document; when you open it, your computer could be taken over without you knowing.

πŸ”§ Prerequisites:

  • Running a supported version of Microsoft Windows or Microsoft Office
  • Opening or processing a maliciously crafted file that utilizes the Windows Codecs Library

⏱ Urgency: High urgency because RCE vulnerabilities in core system libraries like Codecs are frequently exploited in the wild for initial access and lateral movement.

πŸ’‘ Context: The root cause is a flaw in the Windows Codecs Library that allows remote code execution, likely due to improper handling of data during decoding or processing of media files.


CVE-2026-95948 Windows Graphics Kernel Elevation of Privilege Vulnerability

Microsoft MSRC | CVE-2026-95948

❓ Why Should I Care?
Yes, if you run Windows: This is an Elevation of Privilege (EoP) vulnerability in the Windows Graphics Kernel that allows a local attacker to gain SYSTEM-level control of your machine.

🎯 Affected versions: All supported versions of Windows (Windows 10, Windows 11, Windows Server 2016-2025) prior to the specific patch release date.
Not affected: Versions updated with the latest cumulative security patch.

🎭 In plain English:
The core system component that handles graphics and display has a flaw that lets a user with basic access rights escalate to full administrator (SYSTEM) control. An attacker could exploit this to install malware, steal sensitive data, or disable security software without being detected.

πŸ”§ Prerequisites:

  • Attacker must have local access to the Windows system (e.g., via a compromised user account or physical access).
  • The Windows Graphics Kernel must be vulnerable (unpatched).

⏱ Urgency: High urgency for any Windows environment, as local privilege escalation is a critical step in most ransomware and APT attack chains.

βœ… Fixed in: Windows 10/11/Server versions updated with the latest monthly security update (specific build numbers depend on the release date, typically the latest cumulative update).

πŸ’‘ Context: The root cause is likely a memory corruption or logic error in the Windows Graphics Kernel (win32k.sys) that fails to properly validate user input or memory access, allowing privilege escalation.


CVE-2026-95932 Windows Graphics Kernel Elevation of Privilege Vulnerability

Microsoft MSRC | CVE-2026-95932

❓ Why Should I Care?
Yes, if you run Windows 10, 11, or Server 2016/2019/2022: This is a local privilege escalation vulnerability in the Windows Graphics Kernel that allows an attacker with local access to gain SYSTEM privileges.

🎯 Affected versions: Windows 10, Windows 11, Windows Server 2016, Windows Server 2019, Windows Server 2022

🎭 In plain English:
A flaw in how Windows handles graphics commands allows a user with basic access to trick the system into giving them full administrator control. An attacker could exploit this to install malware, steal data, or disable security tools on a machine they already have some access to.

πŸ”§ Prerequisites:

  • Attacker must have local user access to the system
  • The system must be running an affected version of Windows

⏱ Urgency: High urgency for any environment where untrusted users have local access, as this allows trivial escalation to SYSTEM.

βœ… Fixed in: Windows 10 (latest cumulative update), Windows 11 (latest cumulative update), Windows Server 2016 (latest cumulative update), Windows Server 2019 (latest cumulative update), Windows Server 2022 (latest cumulative update)

πŸ’‘ Context: The root cause is a flaw in the Windows Graphics Kernel (gdi32/win32k) that fails to properly validate user-supplied data during graphics operations.


CVE-2026-88217 Windows Graphics Kernel Information Disclosure Vulnerability

Microsoft MSRC

Vendor advisory for an information disclosure vulnerability in Windows Graphics Kernel with no active exploitation reported.

🎯 Affected: Microsoft Windows (Graphics Kernel)

Action: Apply latest Windows security updates


CVE-2026-96150 Microsoft Windows Codecs Library Information Disclosure Vulnerability

Microsoft MSRC

Vendor advisory for an information disclosure vulnerability in Windows Codecs Library with no active exploitation reported.

🎯 Affected: Microsoft Windows (Codecs Library)

Action: Apply latest Windows security updates


CVE-2026-96144 Windows Graphics Kernel Information Disclosure Vulnerability

Microsoft MSRC

Vendor advisory for an information disclosure vulnerability in Windows Graphics Kernel with no active exploitation reported.

🎯 Affected: Microsoft Windows (Graphics Kernel)

Action: Apply latest Windows security updates


CVE-2026-96126 Windows Graphics Kernel Elevation of Privilege Vulnerability

Microsoft MSRC

Vendor advisory for an elevation of privilege vulnerability in Windows Graphics Kernel with no active exploitation reported.

🎯 Affected: Microsoft Windows (Graphics Kernel)

Action: Apply latest Windows security updates


CVE-2026-95953 Windows Graphics Kernel Elevation of Privilege Vulnerability

Microsoft MSRC

Vendor advisory for an elevation of privilege vulnerability in Windows Graphics Kernel with no active exploitation reported.

🎯 Affected: Microsoft Windows (Graphics Kernel)

Action: Apply latest Windows security updates


CVE-2026-95942 Windows Graphics Kernel Elevation of Privilege Vulnerability

Microsoft MSRC

Vendor advisory for a Windows Graphics Kernel Elevation of Privilege vulnerability with no active exploitation reported.

🎯 Affected: Windows Graphics Kernel

Action: Apply latest Windows security updates


CVE-2026-95922 Windows Graphics Kernel Elevation of Privilege Vulnerability

Microsoft MSRC

Vendor advisory for a Windows Graphics Kernel Elevation of Privilege vulnerability with no active exploitation reported.

🎯 Affected: Windows Graphics Kernel

Action: Apply latest Windows security updates


CVE-2026-119861 Windows Internet Connection Sharing (ICS) Elevation of Privilege Vulnerability

Microsoft MSRC

Vendor advisory for a Windows Internet Connection Sharing Elevation of Privilege vulnerability.

🎯 Affected: Windows Internet Connection Sharing (ICS)

Action: Apply latest Windows security updates


CVE-2026-119853 Remote Access API Elevation of Privilege Vulnerability

Microsoft MSRC

Vendor advisory for a Remote Access API Elevation of Privilege vulnerability.

🎯 Affected: Remote Access API

Action: Apply latest Windows security updates


πŸ”΅ On the Radar (15)

  • New Mior and Skyfall Attacks Let Rogue Cloud Hosts Steal Secrets From AMD SEV-SNP VMs (The Hacker News) β€” Two new attacks, Mior and Skyfall, allow a malicious cloud provider or server administrator to steal secrets from virtual machines that are supposed to be encrypted and hidden from the host. Mior can even forge the security reports that prove the VM is running safe software. For example, if you store API keys or customer data in a 'confidential' VM on a public cloud, a rogue host could now read those keys or fake the integrity checks to hide their presence.
  • Check Point gateways hacked via PHP2Shell to deploy VMware ESXi wiper (BleepingComputer) β€” Attackers are actively exploiting critical vulnerabilities in Check Point gateways to wipe data on VMware ESXi servers. This is not a ransomware attack; it is a destructive wiper that deletes data and leaves political notes, indicating a goal of disruption rather than financial gain. If you have unpatched gateways, your infrastructure could be destroyed within hours of exposure.
  • German military drone operations disrupted by cyberattack (BleepingComputer) β€” A cyberattack compromised the update server of a major drone manufacturer, delivering a malicious update to their ground-control software. This forced the grounding of military drones and disrupted operations. For you, this means that if your critical systems (SCADA, IoT, or specialized hardware) rely on vendor update channels, a compromise of that channel can take down your entire operation, not just a single device.
  • Trezor confirms crypto-stealing implants in three wallets, releases fix (BleepingComputer) β€” Trezor confirmed that three Safe 5 devices sold by a specific reseller (CryptoBilis) contained hidden hardware implants that steal recovery phrases and transmit them via cellular networks. If you bought a device from this reseller between August and September 2026, your keys are compromised. Even if you didn't, this highlights a critical supply-chain risk where physical hardware can be tampered with to bypass software security.
  • FBI Says North Korean Hackers Stole Over $400 Million in Crypto Using Deepfake Video Calls (The Hacker News) β€” North Korean hackers are using AI deepfakes and recorded video of trusted colleagues to trick executives and developers into joining fake Zoom/Teams meetings. Once in the call, they claim there is an audio issue and send a link to 'fix' it, which installs malware that steals crypto wallets and hijacks the victim's social accounts to attack their contacts. This is not a theoretical risk; over $400 million has already been stolen, and the attack chain is self-propagating.
  • SilentAir Hijacks Asian Telecom Networks to Intercept SMS 2FA Codes and Spy on Taiwanese Users (The Hacker News) β€” State-level actors have compromised internal equipment at multiple Asian telecom operators to hijack SS7/Diameter signaling, allowing them to divert SMS 2FA codes and spy on specific targets. For infrastructure professionals, this means that SMS is no longer a secure channel for authentication, and the trust model between mobile operators (IPX) is fundamentally broken.
  • Hackers Exploit N-central Flaws to Infect 21 Servers, Including 19 Running Hotfix 2 (The Hacker News) β€” Active exploitation of N-central flaws (4EverAble2Shell, Agent2Shell) has resulted in the compromise of multiple servers, including those with recent hotfixes.
  • PostgreSQL Patches 52 Flaws, Including 22-Year-Old Bug Enabling Authenticated RCE (The Hacker News) β€” Authenticated RCE vulnerability (CVSS 8.8) with a patch available, requiring immediate action to prevent server compromise.
  • Microsoft says November security updates cause printing issues (BleepingComputer) β€” Microsoft's latest security patch introduces a bug where the Windows Graphics Device Interface (GDI) fails to render images that have a zero height or width value. This causes print jobs to fail for documents using these images for watermarks or layout, even if the image is invisible. For example, a user trying to print a contract with a subtle watermark or a complex layout in Word will encounter a print error or blank output after installing the update.
  • How one AI agent exposed five critical enterprise security gaps (BleepingComputer) β€” AI agents can be tricked by hidden text in emails or documents into performing unauthorized actions (like changing bank details) using valid credentials, without triggering standard alerts. For example, an agent reading invoices could be manipulated to update vendor payment info in your ERP, resulting in financial fraud that looks like legitimate user activity in your logs.
  • VMware vDefend Air-Gapped Deployment: Comprehensive Lateral Security for Private Cloud Workloads (VMware / Broadcom Security) β€” Vendor blog post describing a new security deployment model rather than a specific vulnerability or patch.
  • CVE-2026-119396 Windows Remote Access API Tampering Vulnerability (Microsoft MSRC) β€” Informational update to an existing Microsoft advisory with no new technical details or severity changes.
  • Operation Endgame Takes Down Gootloader, CastleRAT and StealC Servers, Seizes 6.4M Stolen Logins (The Hacker News) β€” Law enforcement takedown of known malware infrastructure (Gootloader, CastleRAT, StealC) reduces active threat surface.
  • When an AI Agent Goes Wrong, Can You Stop It? (The Hacker News) β€” Opinion piece discussing the operational challenges of revoking access and stopping AI agents, not a specific vulnerability or breach.
  • Police arrest 354 suspects in dark web cybercrime crackdown (BleepingComputer) β€” Law enforcement action disrupting dark web marketplaces reduces the supply chain for stolen credentials and malware, but requires no immediate technical remediation.

βšͺ 36 low-priority items filtered.


πŸ¦… Aggregated and triaged by Donna AI  |  Sources: 10 vendor feeds  |  CISA KEV

Subscribe to Why Should I Care?

Don’t miss out on the latest issues. Sign up now to get access to the library of members-only issues.
jamie@example.com
Subscribe