CrowdStrike, Google, and The Shadowserver Foundation have disrupted the GlassWorm developer-supply-chain botnet by simultaneously cutting four resilient command-and-control channels. Active since October 2025, GlassWorm spread through malicious OpenVSX and VS Code extensions, GitHub repos, and npm packages (one March campaign hit 400+ artifacts), stealing crypto wallets and developer credentials. Its C2 was built to resist takedown: server addresses encoded in Solana transaction memo fields, configuration stored in the BitTorrent DHT, Base64 C2 paths hidden in Google Calendar event titles, and direct VPS connections for payload delivery. All four had to fall at once. Infected hosts now beacon to CrowdStrike's sinkhole at 164.92.88[.]210.
The FBI has issued a flash alert warning that the Silent Ransom Group (also tracked as Luna Moth, Chatty Spider, and UNC3753) is now sending operatives physically to US law firms to steal data. SRG actors first pose as internal IT over phone or phishing email and try to get an employee to grant a remote-desktop session; if that fails, they dispatch someone in person to plug a USB drive or external hard drive into the target's computer. The group, formed from Conti/BazarCall operators after the 2022 Conti shutdown, has targeted US legal and financial firms since 2023, extorting victims via its leak site.
Israeli firm Gambit Security has forensically linked the late-March attack on the Los Angeles County Metropolitan Transportation Authority to Iran's Ministry of Intelligence and Security (MOIS), despite the attackers branding themselves as the pro-Iran hacktivist collective 'Ababil of Minab.' The group posted videos claiming it wiped hundreds of terabytes and stole over a terabyte of files. LA Metro confirmed the breach on April 2, 2026, and had to check hundreds of servers for compromise before bringing them back online. The case illustrates a recurring pattern of state operations wearing a hacktivist costume to provide deniability while targeting critical infrastructure.
OX Security has flagged a malicious npm package, mouse5212-super-formatter (campaign codenamed Malware-Slop), designed to exfiltrate files from /mnt/user-data - the directory Anthropic's Claude uses to handle uploads and outputs. The package presents itself as an 'archive deployment sync' utility but, during the postinstall stage, authenticates to GitHub using a token found in the victim's environment (or a hard-coded fallback), creates an attacker-controlled repository, and recursively uploads every local file. It writes a fake 'network connections' log to disguise the theft. The package leaked its own GitHub token, suggesting AI-generated malware with poor OPSEC. It has ~676 downloads and remains live on npm.
WatchGuard and ESET have documented two parallel banking-malware campaigns hitting Windows and Android users across Iberia and Latin America. The Windows campaign delivers Grandoreiro - an actively evolving banking trojan operating since 2016 that targets thousands of institutions across 45 countries - via DLL side-loading of four legitimate applications, using Delphi 11-built DLLs that abuse the sgcWebSockets library for WebRTC peer-to-peer C2 over STUN and ICE protocols to blend with web-conferencing traffic. Named targets include Abanca, Banco de Portugal, BBVA PT, Caixa Geral, Santander, plus Revolut and Wise. A companion campaign delivers the BTMOB RAT to Android users in Brazil.
Microsoft has warned of an active cryptojacking campaign that surfaces malicious download sites through AI chatbot recommendations, extending SEO poisoning beyond conventional search. Attackers impersonate legitimate system utilities - CrystalDiskInfo, HWMonitor, Display Driver Uninstaller, FurMark, K-Lite Codec Pack, PDFgear - to target users with high-performance GPUs, prioritizing mining yield per host over mass infection. Beyond mining, the operators deploy ScreenConnect for persistent remote access enabling data theft, lateral movement, or ransomware. Victims who ask LLM-based tools for software-download recommendations are served links to attacker domains on subdomains of gleeze[.]com, hosted via Dynu dynamic DNS. Microsoft says it has detected and blocked the activity.
Microsoft has rolled out a preview of automatic device isolation in Microsoft Defender for Endpoint as part of its automatic attack disruption feature. When Defender detects a compromised endpoint, it now disconnects the device from the network without operator action, while preserving the Defender management channel so the host can still be monitored, investigated, and released. Security teams can release a device from containment after triage via 'Release from isolation' on the Device inventory or device page. The feature works only on onboarded end-user workstations. It joins earlier preview controls for blocking traffic to unmanaged endpoints and isolating compromised user accounts.
Symantec and Carbon Black, working with Huntress, have documented Operation Olalampo, a new MuddyWater (also tracked as Seedworm) espionage campaign that has hit at least nine countries. The Iran-linked actor uses DLL sideloading by abusing two trusted binaries - sentinelmemoryscanner.exe sideloads sentinelagentcore.dll - to deploy the open-source ChromElevator tool, which steals passwords, cookies, and payment-card data from Chromium browsers while bypassing App-Bound Encryption. The campaign also uses Node.js-based implants that drop PowerShell scripts for reconnaissance, SAM-hive theft, screenshot capture, and SOCKS5 reverse-proxy tunneling. Stolen data has been staged on the public file-transfer service sendit[.]sh.
India's national CERT has published a risk-tiered patch SLA blueprint in response to AI-assisted attack acceleration. Internet-facing systems with KEV-listed vulnerabilities must be remediated within 12 hours; critical externally exposed flaws within 1 day; internal-system KEV within 1 day unless documented compensating controls exist; critical internal flaws on high-value systems within 3 days; high-severity issues within 5 days. CERT-In urges defenders to assume breach, adopt Zero Trust and defense-in-depth, embed secure-by-design into AI workflows, validate via red teaming, and treat AI-system visibility as a first-class governance concern. Mitigations (isolation, WAF, monitoring) are expected when no patch is available.
Check Point has documented Iranian APT Nimbus Manticore (also tracked as UNC1549) accelerating its operations during US Operation Epic Fury rather than going quiet. The campaign hits aviation, software, and defense organizations in the US, Europe, and the Middle East via three waves: career-themed phishing using AppDomain hijacking to deploy MiniJunk (February), a trojanized Zoom installer that hijacks legitimate scheduled tasks to deliver the new MiniFast backdoor (March), and the group's first SEO poisoning campaign distributing a weaponized Oracle SQL Developer installer via getsqldeveloper[.]com (April). MiniFast shows signs of AI-assisted development: defensive coding patterns, verbose error strings, and modular structure.
TrustStrike Labs