Urgent Threats & Advisories

Active and archived focus items for SOC teams and threat hunters

Items are AI summaries of the linked sources and are not reviewed before publication. The exception is an item that comes from a rapid alert, which Marcus Lenngren checked and approved. How this works

ACTIVE RIGHT NOW

HIGHADVISORY20h ago

New Windows 11 Narrator Lock Screen 0-Day Allows Attackers to Bypass BitLocker Protections

A zero-day vulnerability in Windows 11's Narrator screen reader allows attackers with physical access to bypass the lock screen and execute code without signing in, even when BitLocker is enabled. The flaw affects Windows 11 version 26H2 with November 2026 security updates. The vulnerability does not break BitLocker encryption or decrypt powered-off drives, but exposes running PCs and allows code execution under a lock screen account.

Action required
Limit physical access to exposed devices and prioritize shutdown or properly configured hibernation over sleep mode until Microsoft releases a FastUnlock-specific advisory.
Windows 11
HIGHADVISORY1d ago

F5 Warns 6 Critical BIG-IP Flaws Can Crash Traffic in Specific Setups, May Allow RCE

F5 disclosed six critical stack-based buffer overflow vulnerabilities in BIG-IP that can crash the Traffic Management Microkernel and may allow remote code execution, but only in systems with specific configurations involving iRules commands or APM policies. The flaws affect BIG-IP, BIG-IP Next for Kubernetes, and BIG-IP CNF across certain versions, and no workaround exists until an engineering hotfix is applied.

Action required
Install the applicable engineering hotfix for your BIG-IP version and configuration: BNK-2.4.2-EHF1, BNK-2.3.4-EHF1, or BNK-2.2.5-EHF1 for BIG-IP Next for Kubernetes; CNF-2.4.2-EHF1, CNF-2.3.4-EHF1, or CNF-2.2.5-EHF1 for BIG-IP CNF.
BIG-IPBIG-IP Next for KubernetesBIG-IP CNFBIG-IP APMBIG-IP DNS
CRITICALADVISORY1d ago

Critical Apache Roller RCE Flaw Under Active Exploitation, No Authentication Required

Apache Roller versions 4.0.0 through 6.1.4 contain a critical remote code execution flaw in the XML-RPC endpoint that allows unauthenticated attackers to execute arbitrary code through insecure deserialization. This flaw is currently being actively exploited in the wild. Any organization running affected versions of Apache Roller is at risk of complete system compromise.

Action required
Upgrade to Apache Roller 6.1.6 or later immediately. If immediate patching is not possible, set enabledForExtensions to false in WEB-INF/web.xml or remove the /roller-services/xmlrpc servlet mapping, then restart Roller.
Apache Roller

ARCHIVE

Category:
Severity:
CRITICALCVE2d ago

nginx2root: unauthenticated HTTP/2 RCE in nginx (CVE-2026-123007), patch to 1.31.9 or 1.30.7 now

A 10-year-old flaw in nginx HTTP/2 flow control lets an unauthenticated attacker run code as the nginx worker. Fixed in 1.31.9 and 1.30.7 on 18 Nov. Chained with a Linux kernel SCTP bug, researchers reached root. PoC is public.

Action required
A reload leaves the old worker running. Fully restart nginx or do a live binary upgrade, or old workers stay vulnerable.
Affected products
nginx 1.31.8
CRITICALCVE3d ago

CVE-2026-114938: Reported pre-auth RCE in Java 17 as Oracle ships 11 Java SE patches

Public reports describe CVE-2026-114938, nicknamed Java4Shell, as an unauthenticated remote code execution flaw in OpenJDK 17.0.6 and later, rated 9.2 under CVSS 4.0. Oracle's November update ships 11 Java SE patches. Details are unconfirmed; here is what to do tonight.

Action required
Running Java 17 does not prove you are exposed. Patch anyway; start with internet-facing services.
Affected products
Java 17 CVE-114938
CRITICALCAMPAIGN3d ago

Microsoft agent-framework 1.26.1 and 1.26.2 on PyPI stole Azure and CI credentials on every Python start

Two malicious releases of Microsoft's agent-framework meta-package landed on PyPI on 17 Nov within 33 minutes. A .pth startup hook ran a credential stealer on every Python launch, targeting Azure, Azure DevOps and GitHub Actions secrets. Pin 1.26.0 and rotate.

Action required
Removing the package does not undo what was sent. Rotate every secret the host could reach, from a clean system.
Affected products
agent-framework
CRITICALEXPLOIT4d ago

ScreenConnect CVE-2026-84869 exploited: malicious files run silently on MSP technician workstations

Attackers are spraying rogue guest sessions at MSP ScreenConnect servers and running malicious batch files on any technician who joins. Patch host clients to 26.6.5 tonight.

Action required
Patching the server is not enough. The flaw is in the client: patch every workstation and Access agent to 26.6.5.
Affected products
ScreenConnect 26.6.5
HIGHADVISORY4d ago

Google Reveals How Hackers Are Using AI Agents to Automate Real-World Cyberattacks

Threat actors are actively using open-source AI agent frameworks (ARTEX, CAIRN, PentAGI, CyberStrikeAI, HexStrike AI) to automate network reconnaissance, privilege escalation, lateral movement, and data theft with minimal human involvement. CAIRN targeted US aerospace organizations, and a custom PentAGI variant scanned organizations across Asia. For the first time, Google found an autonomous 'exploit_agent' framework discovering vulnerabilities and executing exploits without human intervention, including attacks on PaperCut zero-days affecting hundreds of organizations.

Action required
Lock down admin systems, use short-lived credentials, limit access, split networks, monitor for sudden spikes in API and identity activity, set budget alerts for AI services, scan code for leaked keys, and treat any unexpected AI service activation as a serious warning.
Affected products
PaperCutOrganizations across AsiaFrench media, entertainment, and education websitesCrypto firms
CRITICALCAMPAIGN4d ago

OPTIMJACK: Optimizely supply chain attack planted a card skimmer and crypto clipper on 178 brands' sites

On 16 November attackers wrote a loader into the Project JavaScript field of many Optimizely customer projects, so cdn.optimizely.com served a crypto clipper, card skimmer and iOS exploit kits with a valid Optimizely certificate.

Action required
Blocking 1688360.cc is not enough. Some snippets ran an inline Base64 copy. Clean the snippet itself.
Affected products
OPTIMJACK
HIGHADVISORY5d ago

Scans for LiteLLM/Starlette (CVE-2026-42271, CVE-2026-48710), (Mon, Nov 16th)

Attackers exploited CVE-2026-42271 and CVE-2026-48710 in Starlette and LiteLLM over the weekend, targeting honeypots with requests to the /mcp-rest/test/tools/list endpoint. Starlette versions prior to 1.0.1 fail to validate the Host header, allowing unauthenticated attackers to bypass access controls and achieve remote code execution. Attackers are modifying payloads with unusual characters like brackets to evade WAF filters.

Action required
Upgrade Starlette to version 1.0.1 or later.
Affected products
StarletteLiteLLM
HIGHADVISORY6d ago

FBI Says North Korean Hackers Stole Over $400 Million in Crypto Using Deepfake Video Calls

North Korean hackers operating as FaceStealer have stolen over $400 million in cryptocurrency from individuals in the cryptocurrency and Web3 sectors over the past 18 months using deepfake video calls and social engineering. The attack chain begins with compromising a victim's account, then impersonating them to lure their contacts into fake video meetings where malware is deployed. Targets are specifically individuals working in cryptocurrency, Web3, and open-source software development.

Action required
If you work in cryptocurrency, Web3, or open-source software: verify video meeting requests through a separate communication channel before joining, do not click links sent via compromised contacts, and assume any unsolicited meeting request is fraudulent until independently confirmed.
Affected products
ZoomMicrosoft TeamsTelegramLinkedInSlack
HIGHADVISORY6d ago

FBI flash alert says Artex tool automating cyberattacks at ‘machine speed’

The FBI has issued a flash alert about Artex, an open-source AI platform that cybercriminals and nation-states are using to automate cyberattacks at machine speed. The tool can reduce intrusions from initial targeting to privileged access in under three hours and enables actors with minimal expertise to conduct complex operations. Organizations in financial services, government, technology, defense, semiconductor, energy, healthcare, manufacturing, telecommunications and legal industries are being targeted.

Action required
Review your organization's defenses against automated reconnaissance and lateral movement, particularly for third-party integrations, legacy services, and partner portals, which Artex specifically targets.
Affected products
ArtexAnthropic ClaudeOpenAI GPTDeepSeekGLM
HIGHADVISORY7d ago

New SignDrop Exploit Can Abuse Windows Defender to Bypass Security Protections

SignDrop is a Windows privilege escalation zero-day that exploits Microsoft's CLFS authentication mitigation to allow local attackers to gain SYSTEM privileges and disable antimalware defenses. Approximately 80% of Windows installations are vulnerable if they are less than 90 days old or have the mitigation only partially enabled. The flaw can be weaponized against any antivirus or EDR product, not just Windows Defender.

Action required
Check your HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\CLFS\Authentication\Mode registry key; if the value is 0 or 1, your system is vulnerable and you should wait for a Microsoft patch.
Affected products
Windows 11Windows Server 2025Windows Defender
CRITICALEXPLOIT8d ago

Check Point Gaia under attack: PHP2Shell chained to root for ESXi wiper, patch now

Attackers are chaining PHP2Shell (CVE-2026-109618/9) with a DynamicCLI root flaw (CVE-2026-123039) on Check Point Mobile Access gateways, then wiping VMware ESXi hosts. Patch and hunt tonight.

Action required
Blocked-attempt logs are not an all-clear. They do not prove a clean box. Hunt webshells and TSGA directly.
Affected products
Check Point Gaia
CRITICALEXPLOIT8d ago

4EverAble2Shell: N-central Hotfix 2 Servers Taken to Root Hours After PoC Dropped

A new N-able N-central exploit chain, 4EverAble2Shell, gives unauthenticated root on servers running the latest Hotfix 2. Attackers began mining within hours of the PoC, hitting 21 servers, 19 of them patched. Disable Legacy API and take on-prem N-central offline now.

Action required
Hotfix 2 is not enough. Disable Legacy API, take the server off the internet, and treat any hit as full compromise.
Affected products
N-central HF2
HIGHADVISORY9d ago

Grok AI App Builder Exposes Over 35,000 Apps, Leaking Confidential Business Data and Passwords

Researchers found that 94.7% of 37,571 apps built with Grok's app builder were publicly accessible without authentication, exposing sensitive data including patient medical records, employee rosters, plain-text passwords, and internal corporate information. The exposure affected organizations ranging from small contractors to Fortune 500 companies. The vulnerability resulted from overly permissive default access settings and the enumeration of app addresses through passive DNS records.

Action required
Find out which AI app builders your staff use without IT approval, check each app's access settings, test sign-in with an outside account, search the app's script files for passwords, and unpublish any apps that are not actively owned.
Affected products
Grok AI App BuilderLovableReplit
CRITICALCVE9d ago

Citrus ransomware abuses Group Policy to push Sn00py and disable Microsoft Defender across domains

Citrus, a newly documented crew tracked by Securelist, uses stolen contractor logins to reach a domain controller, then builds a malicious GPO named Kaspersky GPO that drops Sn00py ransomware and turns off Defender on every joined machine.

Action required
Defender shows green, so you trust it. A GPO can switch it off by registry. Alert on new GPOs and gPLink changes.
Affected products
Citrus Ransomware
CRITICALADVISORY10d ago

Sparrow framework: China-linked clusters use fake Chrome cert warnings and Linux footholds

Check Point links three China-nexus espionage clusters through the Sparrow framework, run by a supplier it calls CopperDragon. Watering holes push fake Chrome certificate fixes that install the SquadC backdoor, while Linux servers are used as entry points.

Action required
Trusting a cert-repair prompt's branding. Chrome never asks for a fix. Check the download, do not run it.
Affected products
Sparrow / SquadC
CRITICALEXPLOIT11d ago

Microsoft November 2026 Patch Tuesday: exploited Outlook NTLM zero-day among 1,093 fixes

Microsoft's November 2026 Patch Tuesday fixes 1,093 CVEs, including CVE-2026-124194, an actively exploited Outlook spoofing flaw that leaks a user's NTLMv2 hash when a crafted file is opened. Patch Outlook first and block outbound NTLM.

Action required
Treating Moderate as low priority. CVE-2026-124194 is exploited now. Patch Outlook first, block outbound NTLM.
Affected products
CVE-2026-124194
CRITICALEXPLOIT11d ago

Google patches eighth exploited Chrome zero-day of 2026: XML use-after-free CVE-2026-125350

Google has shipped an emergency Chrome update for CVE-2026-125350, a high-severity use-after-free in the XML component that attackers are already exploiting. Update to 157.0.8103.50 and relaunch tonight.

Action required
Background download is not the fix. The patch applies only after you relaunch Chrome. Restart every browser.
Affected products
Chrome 157.0.8103.50
CRITICALADVISORY11d ago

Hackers Used Default admin:admin Credentials to Exploit ForgeRock RCE and Remotely Unlock Cars

ForgeRock Access Management and OpenIDM instances at automotive companies were exploited via default admin:admin credentials combined with URL parsing vulnerabilities, allowing attackers to achieve RCE and access vehicle control APIs. Threat actors gained the ability to unlock vehicles, start engines, and exfiltrate location data. Both affected organizations patched within hours, but similar misconfigurations likely exist elsewhere in the wild.

Action required
Immediately scan your environment for ForgeRock deployments with default credentials enabled. Force credential rotation on all ForgeRock admin accounts and verify no public-facing instances exist. Check ForgeRock logs for authentication attempts using default credentials dating back 90 days.
Affected products
ForgeRock Access ManagementForgeRock OpenIDMApache Felix Web ConsolePing Identity
CRITICALADVISORY12d ago

@ethersphere/bee-js 13.3.0 Compromised

The @ethersphere/bee-js npm package version 13.3.0 was published with malicious code that executes during installation. The payload targets Swarm Desktop wallet data and attempts to establish persistent access across Linux, macOS, and Windows. Any developer or system that installed this specific version is at risk of credential theft and system compromise.

Action required
Immediately identify and isolate any systems with @ethersphere/bee-js 13.3.0 installed. Remove the package, revoke any exposed credentials, and scan for persistence mechanisms. Check package-lock.json and npm audit logs across your environment.
Affected products
@ethersphere/bee-jsSwarm Desktop
CRITICALADVISORY13d ago

Hackers Target OpenSSH SSHole Flaw Within Hours of Disclosure, Gain Root on Honeypot

OpenSSH vulnerability CVE-2026-123158 (SSHole) is under active exploitation within hours of disclosure. Unauthenticated attackers can gain root access on Linux systems with passwordless accounts by sending oversized passwords. Any exposed OpenSSH instance is at immediate risk of full compromise.

Action required
Immediately patch OpenSSH to the latest patched version. Prioritize systems with passwordless accounts or public SSH exposure. Hunt logs for oversized password attempts and failed SSH connections from the past 24 hours.
Affected products
OpenSSHAlpine LinuxPrevidian