GL.iNet Cloud Flaw Exploited to Infect Routers and KVMs
๐จ GL.iNet has confirmed that attackers exploited a flaw in its GoodCloud remote management service to install custom malware on customer routers and KVM devices. Affected users report backdoors and root-level reverse shells. The company has closed the attack path but says this does not remove existing infections, so customers should check devices and apply the latest firmware.
Week in Review: Optimizely Malware Hits 178 Organizations
๐ฐ This week's roundup is led by an attack on Optimizely's web experimentation platform, where malicious JavaScript was planted on customer sites and reached 178 organizations, including BBC, CNN and Zoom. Also in focus is the Java4Shell OpenJDK flaw (CVE-2026-114938), which is tricky to exploit and depends on specific CPUs. Other stories cover five critical NGINX fixes from F5, an exploited Puppet Enterprise bug, and a VMware vCenter zero-day abused for months before a patch.
Bazuka Android Malware Framework Adds LLM Victim Control
๐ก๏ธ ThreatFabric has disclosed Bazuka, a previously unknown Android malware framework with 23 modules covering overlays, keylogging, 2FA and SMS interception, remote control, and crypto-address swapping. Its control panel integrates large language models to help operators score and select victims. A novel overlay technique shows a screenshot of the home or lock screen to hide malicious activity running in the background.
Overlord RAT Hides C2 Addresses on the Solana Blockchain
๐ก๏ธ Cisco Talos has detailed a financially motivated group, tracked as UAT-12643, that hid its command-and-control addresses on the Solana blockchain. The group deployed the Overlord RAT, an infostealer and a clipboard hijacker aimed at cryptocurrency holders, gamers and internet phone users. Overlord resists removal, even crashing Windows if terminated and persisting through a full system reset. Talos assessed with low confidence that the actor is North Korea-aligned.
European Commission Confirms AI-Generated Cyberattack
๐ก๏ธ The European Commission has confirmed that a cyber campaign misusing a publicly available AI tool forced it to take several applications offline. The incident was detected on 22 October and confirmed on 26 October. Most affected systems have since been restored, and no evidence of a staff data breach has been found so far. CERT-EU reportedly linked the activity to the AI-powered tool ARTEX.
ROMULUS ClickFix Platform Delivers CNCMachineRMS RAT
๐จ Researchers at Blackpoint Cyber report that the ROMULUS ClickFix delivery platform is distributing the CNCMachineRMS remote access trojan alongside Legion Loader through separate infection chains. The attacks rely on a malicious MSI installer posing as a Stardock tool and on DLL sideloading to run embedded shellcode. The RAT enables remote desktop control, device capture, and file management. Meanwhile, a wave of other ClickFix campaigns continues to spread.
Attacker Abuses UPS Software to Hit a Domain Controller
๐ Symantec reports that an attacker abused CyberPower's PowerPanel Business software to run malicious scripts on a domain controller. The UPS management tool's script-on-power-event feature runs with SYSTEM privileges, turning a legitimate function into a remote command channel. The intruder staged an antivirus killer, copied the Active Directory database and deployed a persistent implant. The victim and the motive remain unknown.
GTA 6 Scams Target Google Logins, Money, and Crypto Wallets
๐ฎ Scammers are exploiting the GTA 6 launch to steal credentials, collect affiliate commissions, and drain crypto wallets. Malwarebytes researcher Stefan Dasic tracked lures ranging from fake giveaways with cloned Google sign-in pages to code generators that never deliver. A so-called Android app turned out to be an ad-heavy program that cannot be trusted. Players should avoid unofficial offers and never approve unknown wallet requests.
AWS Network Firewall Adds NAT Gateway and Proxy Support
๐ AWS Network Firewall can now be associated with NAT Gateway, working both as an explicit proxy and as a transparent firewall with native network address translation. Both modes are governed by a single, centrally maintained security policy. The feature works with existing capabilities such as managed rule groups, active threat defense, and URL filtering to help block data exfiltration and malware injection.
๐จ Researchers at Calif disclosed Java4Shell (CVE-2026-114938), a critical flaw in the OpenJDK x86-64 Base64 decoder that lets unauthenticated attackers read heap memory, steal secrets, and potentially achieve remote code execution. Exploitability depends on CPUs with AVX-512 VBMI support, which makes standard software inventories unreliable. Fixes are needed from many vendors, not just Oracle, and hidden runtimes complicate remediation for CISOs.
EggStreme APT Retools Malware to Target Philippine Agencies
๐ฅ A suspected Chinese APT group behind the EggStreme framework has rebuilt its toolkit while continuing to target Philippine government bodies, according to Bitdefender. The new agent uses a different network protocol, an expanded command set, and a passive variant that never beacons out. These changes help it evade signatures and heuristics built around the framework after its exposure in September 2025. Related China-aligned activity from Mustang Panda and Earth Alux also hit Southeast Asia.
Chinese APT Rebuilds EggStreme Malware After Bitdefender Report
๐ก๏ธ A suspected Chinese APT group has rebuilt parts of the EggStreme malware framework after Bitdefender exposed it in September 2025. New builds swapped ciphers, moved to raw TCP with mutual TLS, and added Windows domain mapping and in-memory PowerShell. Bitdefender assessed with high confidence that Philippine government organizations were targeted. Core behaviors such as DLL sideloading stayed constant, giving defenders reliable detection points.
How to Spot and Respond to Stalkerware on Your Phone
๐ต๏ธ Kaspersky reports that more than 34,000 people in over 160 countries were affected by stalkerware in 2024-2025, with 33 new families discovered. Its guide explains that monitoring often combines hidden apps, compromised accounts, sharing settings, configuration profiles and physical trackers. Readers are urged to document findings and prioritize personal safety before removing anything, since sudden changes may alert an abuser.
Microsoft Agent Framework PyPI Package Hijacked by Attackers
๐จ Attackers briefly hijacked Microsoft's official Agent Framework package on PyPI, publishing malicious versions 1.26.1 and 1.26.2 that stole cloud tokens, developer secrets and CI/CD credentials. The releases were live for less than an hour but ran automatically whenever Python started. Researchers say the payload targeted Azure workloads and build runners, and teams should rotate any secrets reachable from affected hosts.
UAT-12643 Uses Solana Blockchain for Resilient RAT C2
๐ฐ๏ธ Cisco Talos has identified UAT-12643, a suspected North Korea-aligned actor assessed with low confidence, running a campaign that targets finance staff, VoIP operators, and crypto and gaming users. The group deploys a Go-based Overlord RAT that retrieves its C2 addresses from Solana blockchain memo transactions. Initial access relies on LNK, HTA, and MSBuild XML files that abuse trusted Windows binaries, while persistence even survives a full system reset.
Blockchain Dead Drops Challenge Traditional C2 Defenses
โ๏ธ Talos examines how threat actors are moving command-and-control (C2) configuration data onto public blockchains to escape takedowns. Unlike DGAs, domain fronting, P2P meshes and classic dead drop resolvers, blockchain-hosted records cannot be deleted or altered by any single authority. The report traces the shift from early Bitcoin experiments such as Glupteba to EtherHiding on smart contracts, now used by both cybercriminal and state-sponsored groups.
Japan Takes Fugaku Supercomputer Offline After Cyberattack
๐ Japan's RIKEN Center for Computational Science has taken the Fugaku supercomputer offline after an intruder exploited its public status dashboard to reach internal servers. Between November 10 and 13, the attacker planted malware that targeted user data, including names, email addresses, SSH private keys, and research files. RIKEN has disabled login keys and is rebuilding compromised servers before operations resume.
Custom RATs Target Southeast Asian Governments via HTTP/3
๐ฐ๏ธ Unit 42 has uncovered a cyberespionage campaign, tracked as TGR-STA-0049, that has targeted Southeast Asian government entities since at least July 2025. The actor deployed two undocumented custom RATs, PerfectPioneer and its modular successor KUKURAT. Both rely on covert HTTP/3 (QUIC) command and control to evade inspection. The campaign shows high operational stealth and steady tool refinement over nine months.
CoffeeLoader variant delivered via resurrected Retadup worm
๐ก๏ธ Cybersecurity researchers disclosed a new CoffeeLoader variant that uses a patched 7-Zip DLL as an NSIS installer payload to run malicious shellcode. Kaspersky's analysis shows the loader employs an updated Armoury packer that now leverages OpenGL-based GPU decryption and delivers intermediate shellcode to fetch the main CoffeeLoader payload. The malware establishes persistence via %PROGRAMDATA%/%LOCALAPPDATA% and the Run registry key, prevents autorun disabling, and injects payloads into legitimate processes before contacting hard-coded C2 servers. The loader, linked to Retadup worm infections, continues a campaign that has previously dropped miners, Stop ransomware, and credential stealers.
๐ก๏ธ In 2017 the AutoIt-written Retadup worm spread via removable drives, recruiting infected machines into a botnet. French authorities dismantled its C2 infrastructure in 2019 and issued a self-destruct command, but copies persisted on offline systems and dusty USB sticks. In 2024 attackers re-registered known Retadup domains and relaunched C2 servers, causing dormant worms to retrieve and install the modern CoffeeLoader. CoffeeLoader uses GPU-assisted decryption to evade sandboxes, persists via registry and startup modifications, and downloads additional payloads such as infostealers and ransomware. Kaspersky blocked roughly 15,000 Retadup attempts in the first nine months of 2026, mainly in Peru, Angola and Russia. Recommended precautions include maintaining updated OS and security software, disabling autorun, scanning drives before opening files, revealing extensions, using bootable antivirus tools for old systems, and keeping offline backups.