Multiple npm packages across @tanstack, @mistralai, @uipath, @squawk, and safe-action namespaces were compromised in a worm-like attack affecting over 50 packages. The malicious code executes during installation, downloading the Bun runtime and running a payload that harvests GitHub credentials and cloud secrets. The attack specifically targets AWS environments by querying the IMDS and attempting privilege escalation through STS and SSM endpoints across multiple regions. Stolen credentials are automatically used to publish additional malicious package versions across different maintainer accounts, creating a self-propagating infection chain. The attack patterns mirror previous Shai-Hulud compromises, using a drop-and-execute technique and command-and-control infrastructure at git-tanstack.com, a domain designed to mimic legitimate tanstack.com traffic. Organizations should rotate GitHub credentials, audit AWS credentials, and check for suspicious activity.
Two npm beta releases in the @joyfill namespace were compromised with an import-time JavaScript implant that uses blockchain transactions on Tron, Aptos, and BNB Smart Chain to retrieve encrypted payloads. The malicious code leads to a 77 KB Node.js remote-access trojan identified as DEV#POPPER, which establishes Socket.IO connections for remote control and can execute commands, upload files, read clipboard data, and persist through developer tools. A parallel execution branch downloads additional payloads including an 82 KB Python infostealer assessed to be OmniStealer, targeting browser credentials, Git configurations, and wallet extensions. The compromise affected @joyfill/layouts version 0.1.2-2773.beta.0 and @joyfill/components version 4.0.0-rc24-2773-beta.4, with approximately 16,000 weekly downloads. The loader exhibits exact code overlap with the PolinRider family and DEV#POPPER operations.
A sophisticated ClickFix campaign variant uses social engineering to trick victims into executing commands via the Windows Run dialog. The technique leverages rundll32.exe to load remote non-DLL payloads by ordinal export #1 over WebDAV connections tunneled through HTTPS port 443. Multiple incidents at a single organization show evolving obfuscation methods including WMI process spawning, caret insertion, and runtime string assembly to evade detection. The attack chain utilizes trusted Windows binaries like pcalua.exe to break process lineage tracking. No files are dropped to disk, and payloads are invoked by ordinal rather than named functions. Successful attacks exfiltrated browser credentials and sensitive documents totaling 13MB. The most obfuscated variant evaded automated EDR detection entirely, being discovered only through proactive threat hunting focused on ordinal execution patterns rather than keyword detection.
Researchers uncovered a fractured criminal ecosystem built around Flying Eagle, an Android remote access tool whose source code was stolen in early 2026 along with nearly 200 customer databases. The investigation began with a malicious APK impersonating a Chinese Provincial Public Security Bureau app, leading to identification of 170 active servers running the framework. Two Telegram channels, SQLRCE0 and Yx Technology, distribute modified versions with operational support and cash-out services at 20-50% transaction fees. The platform combines APK generation and C2 device management with phishing overlays targeting financial, adult, and government services. A successor platform called Night Dragon was introduced in June 2026, featuring enhanced credential capture capabilities for Chinese banking apps, cryptocurrency wallets, and social media platforms. The activity primarily targets Chinese citizens through social engineering lures, though international templates suggest broader targeting capability.
Between May and July 2026, security researchers deployed an unauthenticated Model Context Protocol (MCP) honeypot server to observe how threat actors exploit AI agent infrastructure. Of approximately 1,000 sources that reached the decoy, 596 spoke the protocol and 24 proceeded to actively exploit it. These operators executed 628 shell commands, 255 file reads, and 248 secrets-store lookups, with 19 hunting credentials and 4 attempting container escapes. Activity escalated from 39 tool calls in May to 877 by mid-July. Three stolen credentials were subsequently used against a live AWS account, with two cases involving Bedrock model invocation for LLMjacking. The attacks demonstrated automated reconnaissance, credential harvesting, container escape attempts, backdoor account creation, and Kubernetes enumeration, revealing that exposed MCP servers represent a growing attack surface as AI agent infrastructure proliferates.
Since January 2026, a threat actor likely functioning as an initial access broker for ransomware operations has been targeting organizations through Microsoft Teams vishing attacks. Attackers impersonate IT helpdesk staff to convince victims to initiate Quick Assist remote sessions. Following initial compromise, PowerShell scripts deploy a Go-based backdoor called GoGRPC, which exists in four distinct variants: Lep, Giver, Pet, and Kind. These variants communicate with command-and-control infrastructure using gRPC over HTTP/2, an uncommon approach that helps blend malicious traffic with legitimate communications. Additional tools observed include BlindDoor backdoor, RevSocket and PyGRPC SOCKS proxies, S3Siphon data exfiltration utility, and RSOX Rust-based proxy relay. Recent campaigns show increased sophistication and selectivity, with heightened focus on corporate environments through enhanced PowerShell scripts capable of antivirus detection, domain controller fingerprinting, and system reconnaissance b...
A fraudulent website impersonating Corepack, the Node.js package manager tool, is distributing malware to developers. The attackers exploit timing around Corepack's removal from Node.js bundling, targeting developers searching for installation instructions. The site offers Windows executables that deliver OpenShield infostealer and proxyware, enrolling victim machines in bandwidth-sharing networks without consent. The payload steals browser credentials, SSH keys, establishes persistence, and routes third-party traffic through compromised systems. An alternative download path delivers adware and trojan components disguised as OperaGX installer. The site features AI-generated content with obvious errors, including confusing Yarn package manager with textile crafts. The domain has been reported to registrars for takedown after community members identified the threat.
AhnLab monitored Advanced Persistent Threat attacks targeting South Korea during June 2026, identifying multiple attack types distributed primarily through spear phishing campaigns. Threat actors disguised malicious files as work-related documents, with LNK files being the most common delivery method. Six distinct attack types were observed, employing various techniques including malicious PowerShell commands, AutoIt malware, curl.exe abuse, GitHub repository exploitation, Task Scheduler persistence, DLL side-loading, and Python backdoors. These attacks deployed Infostealers, keyloggers, backdoors, and remote access tools like XenoRAT. Once executed, the malware established persistence, exfiltrated system information, and enabled remote control of compromised systems. Organizations are advised to verify email senders, avoid opening files from unknown sources, apply security patches, and maintain updated antivirus software to mitigate these persistent threats.
The Intercom TypeScript Library version 7.0.4 has been compromised with malicious code that harvests GitHub credentials. Upon installation, the package executes a preinstall hook that downloads the Bun runtime, then runs a payload to extract GitHub credentials using the gh auth token command. The attack employs sophisticated C2 communication by querying GitHub's commit search API for specific strings embedded in public repositories, effectively using legitimate services to evade detection. The attack patterns mirror previous Shai-Hulud compromises, which exhibit worm-like behavior by automatically using stolen credentials to infect additional npm packages. With 361,510 weekly downloads, this compromise poses significant risk for a widespread infection wave similar to November 2025 when over 1,000 packages were affected.
Insikt Group identified four new malware families from TAG-195 (Golden Chickens, Venom Spider), a financially motivated malware-as-a-service developer. The families include TinyEgg, a lightweight initial-access backdoor; ChonkyChicken, which expands capabilities with browser credential theft and session automation; a modularized ChonkyChicken variant using controller-and-plugin architecture; and ChromEggscalator, a modified Chrome encryption-bypass tool. TAG-127 has been observed deploying TinyEgg via ClickFix campaigns using fake security verification pages. The modular architecture reduces static detection exposure and enables selective capability provisioning to operators. All families share consistent architectural traits including WebSocket command-and-control, Run key persistence, string obfuscation, and execution via legitimate Windows binaries. This represents a deliberate architectural transition toward operator-driven tooling within the TAG-195 MaaS ecosystem.
During Q2 2026, Microsoft detected approximately 7.6 billion email-based phishing threats, with monthly volumes declining from 2.7 billion in April to 2.4 billion in June. The quarter was significantly shaped by the downstream effects of Microsoft's Digital Crimes Unit disruption of the Tycoon2FA phishing-as-a-service platform in March, resulting in a 92% decline in associated phishing volume. QR code phishing attacks peaked at 18.7 monthly attacks in March before declining 48% by June, while CAPTCHA-gated phishing fell 81% from its March high. Credential phishing remained the dominant objective, accounting for 94-96% of all payload-based attacks. Business email compromise activity returned to historical norms after an anomalous April surge. Microsoft Teams-based threats grew substantially, with weekly malicious call attempts reaching nearly ten times the mid-2025 baseline by quarter end.
Helix is a data extortion group conducting multi-target campaigns using vishing, device code phishing, and automated SharePoint data exfiltration. Emerging from the BlackFile and ShinyHunters ecosystem after BlackFile's shutdown in April 2026, Helix uses social engineering to impersonate managers during vishing calls to trigger device code authentication flows. After compromising accounts, the group registers MFA to maintain access, enumerates SharePoint sites with automated tools, and exfiltrates bulk data. Their infrastructure shares hosting with BlackFile and uses geo-matched residential proxies to evade detection. The group varies dwell times to complicate incident response efforts.
Device code phishing exploits the OAuth 2.0 device authorization grant, a legitimate authentication feature designed for input-limited devices like smart TVs. Attackers initiate a device-code request with Microsoft, receive a valid code, then trick victims into approving it through social engineering. The victim authenticates on genuine Microsoft pages and completes MFA, but the session tokens are issued to the attacker instead. When targeting the Microsoft Authentication Broker, attackers can register rogue devices and obtain long-lived refresh tokens for persistent access. A recent campaign used sophisticated multi-stage delivery chains involving Google Sites, compromised website redirectors, and fake document-sharing portals. After successful authentication, attackers registered multiple devices, created hidden mailbox rules, and used compromised accounts to send additional phishing emails, all without touching victim endpoints.
A malvertising campaign targeted 29 organizations by exploiting the legitimate Claude.ai domain to distribute a fake ClaudeDesktop.exe containing the SectopRAT remote access trojan. The campaign used advanced anti-analysis techniques such as VMProtect packing, GPU-based VM detection, and DirectX shader-based payload decryption. The malware's command-and-control infrastructure leveraged Ethereum blockchain transactions (EtherHiding) to hide C2 addresses, enhancing resilience against takedowns. SectopRAT exfiltrated sensitive data including credit card information, credentials, browser data, and personal files. The campaign infrastructure is linked to previous operations dating back to December 2025 and associated with known threat groups.
A newly discovered Windows stealer and remote access trojan called Dolphin X targets over 300 applications including browsers, cryptocurrency wallets, password managers, and cloud command-line tools. The malware collects credentials ranging from browser logins to SSH keys, .env files, and cloud tokens. A distinctive feature called the AI Profiler automatically scores infected victims based on application usage, browsing activity, and installed software, helping attackers identify high-value targets through daily summaries. The malware builder operates through a remote server that compiles agents and offers optional mutation engines to evade detection. Advertised by a vendor using the alias Kontraktnik, Dolphin X poses significant risk to developers and organizations by potentially exposing access to entire production environments through compromised DevOps credentials.
A TrickBot variant has been identified that uses DNS tunneling for command-and-control communications instead of traditional HTTP protocols. The malware maintains persistence through Windows Task Scheduler, creating disguised tasks that execute at startup and repeat every five minutes. Configuration data is stored in NTFS Alternate Data Streams to evade detection. The malware employs multiple obfuscation techniques including encrypted strings, runtime API resolution via hash-based lookups, and dynamically calculated constants. Its modular architecture supports twelve different control commands enabling capabilities such as module downloads, process injection through hollowing and doppelgänging techniques, PowerShell execution, and raw machine code execution. The variant transfers data through specially crafted DNS queries to public DNS servers, encoding command data in malformed domain names and receiving responses embedded within multiple IPv4 addresses, achieving transfer speeds of approximately 30.7 KB/s.
A sophisticated phishing campaign impersonates legitimate business entities including UPS and the Malaysian Inland Revenue Board to distribute Phantom Stealer v3.5.0. The attack begins with convincing emails containing compressed archives housing malicious JavaScript files. Once executed, the JavaScript launches obfuscated PowerShell scripts that operate entirely in memory, deploying multiple stages of encrypted and encoded payloads. The infection chain utilizes Base64 encoding, AES encryption, and XOR ciphering to conceal its activities. The final payload, Phantom Stealer, harvests credentials from browsers, cryptocurrency wallets, messaging applications, and system information before exfiltrating stolen data via SMTP over port 587 using STARTTLS encryption. The multi-layered approach significantly reduces on-disk footprint and employs reflective code loading and process injection into legitimate binaries to evade traditional security defenses.
A sophisticated adversary-in-the-middle phishing campaign is targeting universities, enterprises, and multinational institutions including EU and UN agencies. Active since May 2026, the operation leverages compromised organizational email accounts to distribute procurement-themed lures that mimic requests for information, bid invitations, and project documentation. Victims are redirected through fake document portals, CAPTCHA verification stages, and cloned authentication pages impersonating Microsoft, OpenGov, and financial institutions. The attacker rotates between multiple AiTM phishing kits including EvilProxy, FlowerStorm, and Kali365 to intercept MFA-protected sessions in real time, capturing session tokens and cookies to establish authenticated access. Rather than using newly registered domains, the actor compromises aged domains that have been dormant for years, injecting PHP files to host phishing infrastructure and evade detection through domain reputation systems.
An active Lampion malware campaign has been identified targeting Portuguese users through phishing emails that impersonate financial and administrative communications. Lampion, a Brazilian banking malware derived from the ChePro lineage, delivers initial payloads via ZIP archives containing heavily obfuscated HTML files designed to evade detection. The HTML stage retrieves additional scripts from attacker-controlled infrastructure, initiating a multistage VBS infection chain. Each stage employs extensive obfuscation techniques including junk code, encrypted strings, and dynamically generated scripts that inflate file sizes while concealing core functionality. The infection chain is deliberately fragmented across multiple independent execution stages, complicating behavioral analysis. Telemetry shows 94.6% of detections concentrated in Portugal, confirming this is a highly targeted threat focused on Portuguese-speaking victims.
Kaspersky identified a new .NET Native AOT communication module for Project CAV3RN, a sophisticated cyberespionage framework targeting Israel. This module replaces the previous HTTP/WebSocket component by exchanging commands through Outlook calendar events accessed via Microsoft Graph API. Commands and results are stored in a fixed time window (2050-05-13 22:00-23:00 UTC) using specific subject patterns to identify heartbeats, commands, and outputs. The module implements RSA and AES-GCM encryption for secure communications. If Microsoft Graph authentication fails, it retrieves replacement configuration settings through DNS AAAA record responses from actor-controlled nameservers. The infrastructure analysis and behavioral patterns suggest low-confidence attribution to OilRig (APT34), based on the use of Microsoft-hosted services for command and control and compromised regional infrastructure.
