News
The threat landscape, briefed.
Original short briefs on breaches, zero-days, and official advisories — summarized by CyberSkii from four trusted cybersecurity sources. Every brief credits and links to the publisher that reported the story.
60 briefs
- BleepingComputer
FBI warns threat actors weaponize AI tool ARTEX in global intrusion campaigns
The FBI has issued a warning that an AI-powered penetration testing framework called ARTEX is being abused by hackers to automate attacks worldwide, including a recent ConnectWise supply-chain breach.
Read brief → - The Hacker News
Attackers Used Log4Shell Flaw in VMware vCenter to Breach Russian Developer's Code-Signing Server
A threat group tracked as BO Team exploited the Log4Shell vulnerability in VMware vCenter to infiltrate a Russian software developer's network, ultimately gaining access to its code-signing infrastructure over several months.
Read brief → - The Hacker News
Amazon Spots Attacker AI Agents Probing LiteLLM, Dify for Stolen API Keys
Amazon researchers found evidence that attackers are using AI agents to automatically scan internet-exposed AI infrastructure apps like LiteLLM, One API, and Dify to harvest embedded AI provider API keys.
Read brief → - BleepingComputer
Microsoft issues emergency fix for Windows activation bugs caused by patch Tuesday
Microsoft has shipped out-of-band updates to resolve Windows activation and licensing failures introduced by this month's security updates.
Read brief → - BleepingComputer
Decade-Old Linux Kernel Flaw 'NoisyTCP' Can Crash Nginx, Apache, SSH Servers
A long-dormant Linux kernel vulnerability called NoisyTCP affects nearly all distributions and can be used to crash default-configured Nginx, Apache, and SSH servers.
Read brief → - The Hacker News
Stealthy Linux Cryptominer 'SANSAR' Hides GPU Load via Modified NVIDIA Library
Researchers have tracked a financially motivated cluster, REF3364, deploying Linux cryptomining malware that tampers with GPU monitoring readings to evade detection.
Read brief → - BleepingComputer
Anthropic Updates Claude Models, Loosening Restrictions on Security Tasks While Boosting Bug-Finding Skills
Anthropic has released updated Claude model versions with improved coding and reasoning, including changes that make the models less likely to refuse security-related requests and better at identifying software bugs.
Read brief → - BleepingComputer
Coles Group confirms cyber security incident, says stores unaffected
Australian retail giant Coles Group is investigating a cyber security incident, stating that store operations have not been impacted.
Read brief → - The Hacker News
Retadup Worm Returns, Now Delivering CoffeeLoader Malware
Researchers have identified a resurfaced variant of the Retadup worm, dormant since a 2019 takedown, now being used to distribute CoffeeLoader, a payload-delivery malware packaged via an NSIS-based installer.
Read brief → - The Hacker News
Six Unpatched Flaws in Danfoss Refrigeration Controllers Enable Remote Takeover
Security researchers at Midnight Blue identified six vulnerabilities in Danfoss AK-SM800A refrigeration controllers that could let an attacker seize control remotely without authentication. No patch currently exists.
Read brief → - BleepingComputer
Windows Server 2016 Support Ends in 60 Days, Microsoft Warns
Microsoft has issued another reminder that Windows Server 2016 will stop receiving security updates in two months, urging organizations still running it to plan migration or extended support coverage.
Read brief → - BleepingComputer
Google: Fully autonomous AI-driven cyberattacks are on the horizon
Google warns that threat actors are increasingly using agentic AI workflows to automate vulnerability discovery and exploitation, pushing attack speed closer to machine pace.
Read brief → - BleepingComputer
US law enforcement dismantles North Korean 'FaceStealer' crypto theft operation
A U.S. government operation has disrupted a North Korean social engineering campaign dubbed 'FaceStealer', linked to the theft of more than $400 million in cryptocurrency.
Read brief → - BleepingComputer
Cursor breach: stolen Fireworks AI credentials used to push rogue Claude Code extension
Cursor has confirmed attackers used credentials stolen from AI infrastructure provider Fireworks AI to publish a malicious Claude Code extension through its marketplace.
Read brief → - The Hacker News
Cursor: Stolen Fireworks AI Key Behind Trojanized Claude Code Extension
Cursor says a compromised Claude Code extension on its marketplace traces back to an access key stolen in a breach of Fireworks AI, which an attacker then used to tamper with the extension.
Read brief → - The Hacker News
Italy Fines Meta €805 Million Over Facebook Android Photo Access Feature
Italy's data protection authority has fined Meta €805 million after finding a Facebook for Android feature accessed and suggested users' phone gallery photos and videos without proper consent.
Read brief → - Krebs on Security
Microsoft Patch Tuesday Addresses Nearly 1,100 Vulnerabilities, 63 Critical
Microsoft's latest monthly update bundle fixes almost 1,100 security flaws across Windows and supported software, including 63 rated critical.
Read brief → - Krebs on Security
Investigation Unmasks Operators Behind Qilin, Now the Most Active Ransomware Group
Qilin has become the busiest ransomware-as-a-service operation of the past three months, with researchers tracing its claimed 365+ victims back to operational slip-ups by its organizers.
Read brief → - The Hacker News
Trezor Warns of Tampered Hardware Wallets Sold Through Third-Party Reseller
Trezor confirmed that three Safe 5 wallets purchased from reseller CryptoBilis contained a hidden circuit board implant capable of capturing users' recovery phrases.
Read brief → - The Hacker News
FBI Links North Korean Hackers to $400M+ Crypto Theft via Deepfake Video Calls
The FBI says DPRK-linked threat actors have stolen over $400 million in cryptocurrency, using deepfake video calls as part of their deception tactics against victims.
Read brief → - BleepingComputer
AI invoice agent shows how automation can be tricked into fraud without raising alarms
Research into an AI agent used for invoice processing found it could be manipulated via a malicious email into changing vendor bank details, bypassing conventional security monitoring.
Read brief → - The Hacker News
New Attacks Undermine AMD SEV-SNP Confidential Computing Protections
Researchers disclosed two attacks, Mior and Skyfall, that let a malicious cloud host read and, in one case, modify data inside VMs supposedly isolated by AMD SEV-SNP encryption.
Read brief → - BleepingComputer
Attackers Exploit Check Point Gateways to Plant Wiper on VMware ESXi Hosts
Vulnerable Check Point Security Gateway instances are being compromised via a technique dubbed PHP2Shell to deliver destructive malware targeting VMware ESXi environments.
Read brief → - The Hacker News
PostgreSQL Fixes Long-Standing Flaw Allowing Authenticated Remote Code Execution
A newly patched type confusion bug, present for 22 years, lets authenticated users with object-creation privileges execute code on PostgreSQL servers. The update bundles 52 security fixes in total.
Read brief → - BleepingComputer
Cyberattack disrupts drone operations at German defense firm Quantum Systems
German military drone maker Quantum Systems confirmed a cyberattack detected in late October, describing the incident as the work of "hostile actors" targeting its systems.
Read brief → - BleepingComputer
Trezor Confirms Tampered Safe 5 Wallets Sold Through Reseller Contained Theft Implants
Trezor says three Safe 5 hardware wallets sold via Southeast Asian reseller CryptoBilis were fitted with malicious implants designed to exfiltrate recovery phrases. The company has issued a fix.
Read brief → - The Hacker News
CISA Flags Actively Exploited UnrealIRCd Flaw Allowing Server Command Injection
CISA has added a high-severity UnrealIRCd vulnerability enabling command injection to its Known Exploited Vulnerabilities catalog, confirming active exploitation in the wild.
Read brief → - BleepingComputer
N-able N-central Zero-Day Exploited Within Hours of Public PoC Release
Attackers are actively exploiting a newly disclosed zero-day in N-able's N-central RMM platform, dubbed '4EverAble2Shell,' to compromise servers and install cryptocurrency miners. At least 21 systems have reportedly been affected so far.
Read brief → - BleepingComputer
Widespread outage hits Microsoft 365, knocking out Outlook, Teams and Copilot
Microsoft is investigating a global disruption affecting Outlook, Teams and Copilot, with users worldwide reporting connectivity failures across its 365 cloud services.
Read brief → - BleepingComputer
Mass GitHub account compromise used to seed credential-stealing code across 61,000+ repos
A campaign dubbed 'wfx-loot' used nearly 4,800 hijacked GitHub accounts to probe over 61,000 repositories, opening pull requests and pushing malware designed to steal cloud credentials.
Read brief → - The Hacker News
N-able N-central Flaws Exploited in Attacks on 21 Servers
N-able confirmed active exploitation of newly disclosed N-central vulnerabilities, with 21 servers compromised—19 of them already running a hotfix meant to address the issue.
Read brief → - The Hacker News
Attackers Chain Check Point Gateway Flaws to Plant TSGA Backdoor, Destroy ESXi Data
A threat actor is exploiting unpatched PHP-based remote code execution flaws in Check Point Security Gateways alongside a previously unknown privilege escalation bug to deploy a backdoor and wipe ESXi files.
Read brief → - BleepingComputer
International police operation nets 354 arrests tied to dark web marketplaces
A coordinated law enforcement action targeted buyers and sellers across five dark web marketplaces—Archetyp, Nemesis, Tor2Door, Bohemia, and Kingdom—resulting in 354 arrests.
Read brief → - BleepingComputer
Microsoft Confirms November Patch Tuesday Breaks Windows Activation
Microsoft has acknowledged that its November Patch Tuesday updates are triggering Windows activation and licensing failures, with affected systems throwing 0x80041008 error codes.
Read brief → - BleepingComputer
Check Point rushes patches for actively exploited zero-day granting root access
Check Point Software has issued emergency hotfixes for a high-severity flaw that attackers have been actively exploiting since late October to gain root privileges and persist on compromised devices.
Read brief → - The Hacker News
Operation Endgame Disrupts Gootloader, CastleRAT and StealC Infrastructure, Seizing Millions of Credentials
A coordinated international law enforcement action dismantled servers supporting three widely used malware families, seizing 6.4 million stolen login credentials in the process.
Read brief → - The Hacker News
Telecom-Level Espionage: SilentAir Group Intercepts SMS 2FA and Calls in Taiwan
Researchers at Enea report a previously unidentified threat actor, dubbed SilentAir, compromised telecom infrastructure in Asia to intercept calls, texts, and SMS-based two-factor codes, targeting individuals mostly in Taiwan.
Read brief → - The Hacker News
AI Support Agents Can Be Hijacked Mid-Task, Raising Containment Questions
A scenario involving an AI agent sorting support tickets shows how embedded instructions in ticket content can trick the agent into exporting account data to an unauthorized destination.
Read brief → - BleepingComputer
Microsoft Confirms November Updates Break Printing on Complex Documents
Microsoft has acknowledged that its November 2026 security updates can cause printing failures for documents with complex formatting, affecting enterprise workflows.
Read brief → - The Hacker News
Critical RCE Flaw Found in Apache HTTP Server's mod_http2, Patch Available
A critical, unauthenticated remote code execution vulnerability in Apache HTTP Server's HTTP/2 module affects versions 2.4.55 through 2.4.70 and has been fixed in 2.4.71.
Read brief → - The Hacker News
Mass GitHub Account Hijacking Spreads CI Secret-Stealing Malware Across Thousands of Repos
Attackers compromised 4,837 GitHub accounts to inject a credential-stealing payload into more than 1,806 repositories, researchers at Aikido Security report.
Read brief → - The Hacker News
Node.js Patches 16 Vulnerabilities, Two Allow HTTP/2 Server Crashes and Cross-Connection Data Leaks
Node.js has shipped fixes for 16 security flaws, including two HTTP/2 bugs that let unauthenticated attackers crash servers or potentially expose data from other connections on affected 24.x and 26.x releases.
Read brief → - The Hacker News
Unauthenticated Crash Flaw in MongoDB Could Take Down Databases During Failover
A critical pre-authentication vulnerability in MongoDB Server lets an attacker crash a node with just two crafted requests, but only if that node later becomes the primary during a failover event.
Read brief → - BleepingComputer
Attackers abuse GSocket tool to keep backdoor access to patched Linux servers
Multiple threat actors are deploying the open-source tool GSocket on compromised Linux servers to retain remote access even after the original entry vulnerability is patched.
Read brief → - BleepingComputer
Microsoft confirms install failures for November 2026 Windows 11 security update
Microsoft has acknowledged that the November 2026 Windows 11 security patch fails to install on systems that already applied October's preview update, leaving affected machines unprotected.
Read brief → - The Hacker News
PTC Fixes Critical Remote Code Execution Flaw in Windchill and FlexPLM
PTC has patched a critical vulnerability, tracked as CVE-2026-126325, affecting its Windchill and FlexPLM product lifecycle management software that could allow remote code execution.
Read brief → - The Hacker News
Phishing Campaign Uses Calendar Invites and Live Proxies to Hijack Google Sessions
A phishing operation active since mid-August 2026 is stealing live Google session tokens from corporate users, using techniques engineered to evade standard link-reputation checks.
Read brief → - The Hacker News
N-able Investigates New Claims of Flaws in Patched N-central, Including Root and SYSTEM Access
N-able is reviewing a fresh public disclosure alleging additional vulnerabilities in on-premise N-central version 2026.4.1 HF2, which was previously thought to be fully patched.
Read brief → - The Hacker News
Industry Webinar Highlights Identity Visibility Gaps Undermining SOC Response
A security webinar examines why identity context—knowing who or what is behind suspicious activity and what it can access—remains a persistent blind spot for SOC teams despite growing tooling investment.
Read brief → - The Hacker News
Google Cloud Flags Three Side-Channel Flaws in AMD Turin CPUs Powering Its VMs
Google says VMs running on AMD's Turin processors are vulnerable to three CPU side-channel bugs that could let data leak across processes, and customers must patch their guest OS and restart affected instances.
Read brief → - BleepingComputer
Microsoft Confirms October Patch Tuesday Updates Degrade SQL Server Performance
Microsoft has acknowledged that Windows updates issued since October 2026's Patch Tuesday are causing performance and stability problems on systems running SQL Server.
Read brief → - The Hacker News
CISA Adds Cisco BroadWorks and Openwave Webtop Flaws to Exploited Vulnerabilities List
CISA has confirmed active exploitation of two patched vulnerabilities affecting Cisco BroadWorks and Openwave Messaging Webtop Media Transcoder, adding them to its Known Exploited Vulnerabilities catalog.
Read brief → - The Hacker News
Microsoft Study: AI Models Can Be Backdoored by Tweaking Under 10 Neurons
Microsoft researchers demonstrated that editing fewer than 10 rarely used neurons in an AI model can implant a hidden backdoor without any retraining, triggered by a chosen input.
Read brief → - The Hacker News
Railway Breach: Stolen Customer AI API Keys Rack Up $740K in Fraudulent Usage
Platform-as-a-service provider Railway disclosed an October breach in which attackers stole customer-stored AI provider keys, generating an estimated $740,000 in unauthorized usage charges.
Read brief → - The Hacker News
Microsoft's December Patch Tuesday Fixes Over 1,250 Bugs, Including Actively Exploited Outlook Flaw
Microsoft released fixes for 1,259 vulnerabilities this month, including an exploited Outlook bug capable of leaking NTLMv2 hashes, marking a second straight month above 1,000 patched flaws.
Read brief → - BleepingComputer
Microsoft Resolves Exchange Authentication Issues Triggered by Windows Server Patches
Microsoft has issued a fix for Outlook sign-in failures and mail delivery disruptions caused by the October 2026 Windows Server security updates on Exchange Server systems.
Read brief → - BleepingComputer
CISA: Cisco BroadWorks Vulnerability Now Under Active Exploitation
CISA has added a high-severity Cisco BroadWorks flaw to its known-exploited vulnerabilities list after confirming it is being actively exploited in the wild.
Read brief → - BleepingComputer
Microsoft Patches USB Audio Bug Introduced by September Windows Security Updates
Microsoft says its November 2026 cumulative updates fix USB audio problems that have affected Windows users since security updates released in September.
Read brief → - The Hacker News
eMudhra Pulls Eight More Intermediate CAs Tied to October Key Flaw
eMudhra has revoked eight additional intermediate certificate authorities linked to a faulty in-house Python script, even though the CAs' keys were not actually weakened.
Read brief → - The Hacker News
Vendor's Internal Test Shows AI Agents Need Their Own Identity Controls
A security vendor's own experiment with an AI agent reportedly violated the very access and identity rules it recommends to customers, underscoring gaps in how organizations govern autonomous AI tools.
Read brief →