A persistent nine-month RondoDoX botnet campaign has been targeting IoT devices and web applications. The threat actors have recently shifted to weaponizing a critical Next.js vulnerability, deploying malicious payloads like 'React2Shell' and cryptominers. The campaign, spanning from March to December 2025, shows quick adaptation to latest attack trends. The activity is divided into three phases: initial reconnaissance, web application exploitation, and IoT botnet deployment. The attackers have been using multiple command and control servers and deploying various malware variants. The campaign has intensified in December 2025 with a focus on Next.js exploitation. The impact includes widespread IoT device compromise, Next.js application risks, credential harvesting, and persistent multi-architecture threats.
In mid-2025, a malicious driver file was discovered on Asian computer systems, signed with a compromised digital certificate. This driver injects a backdoor Trojan and protects malicious files, processes, and registry keys. The final payload is a new variant of the ToneShell backdoor, associated with the HoneyMyte APT group. The attacks, which began in February 2025, primarily target government organizations in Southeast and East Asia, especially Myanmar and Thailand. The malware uses various techniques to evade detection, including API obfuscation, process protection, and registry key protection. The ToneShell backdoor communicates with command-and-control servers using fake TLS headers and supports remote operations such as file transfer and shell access.
A sophisticated campaign by the Chinese APT group Silver Fox is targeting Indian entities with authentic-looking Income Tax phishing lures. The attack leverages a complex kill chain involving DLL hijacking and the modular Valley RAT to ensure persistence. The campaign uses a multi-stage infection process, starting with a malicious email containing a PDF decoy. The payload is delivered through an NSIS installer, which drops a legitimate Thunder.exe binary and a malicious libexpat.dll for DLL hijacking. The final stage involves the Valley RAT, which uses a two-stage configuration loading mechanism and implements a 3-tier C2 communication loop. The RAT's modular plugin architecture allows for dynamic capability extension and persistence through registry-based storage.
The Evasive Panda APT group conducted highly-targeted campaigns from November 2022 to November 2024, employing adversary-in-the-middle attacks and DNS poisoning techniques. They developed a new loader that evades detection and uses hybrid encryption for victim-specific implants. The group utilized fake updaters for popular applications to deliver malware, including a multi-stage shellcode execution process. A secondary loader, disguised as a legitimate Windows library, was used to achieve stealthier loading. The attackers employed a custom hybrid encryption method combining DPAPI and RC5 to secure payloads. Victims were detected in Türkiye, China, and India, with some systems compromised for over a year. The campaign showcases the group's advanced capabilities and continuous improvement of tactics.
During the holiday season, threat actors exploit overloaded inboxes and financial stress through two main patterns: Docusign-themed phishing for corporate credential harvesting and loan offer spam for personal data theft. The Docusign campaign uses spoofed emails with authentic-looking branding, redirecting through disposable hosting platforms to a credential harvesting page. The loan scams range from obvious 'Xmas loan' offers to sophisticated marketing-style emails, ultimately leading victims to a detailed identity theft questionnaire on christmasscheercash.com. Both scams utilize seasonal themes and mimic normal end-of-year workflows to increase effectiveness. Defensive measures include verifying sender domains, validating link destinations, and treating unsolicited loan offers as high risk.
This analysis challenges the notion that cyber threat actors are always sophisticated and organized. Through examining three incidents, it reveals that attackers often make mistakes, face obstacles, and adapt their tactics based on trial and error. The incidents showcase how threat actors struggled with Windows Defender, mistyped commands, and failed to start malicious services. Despite using similar tactics and infrastructure across attacks, the perpetrators had to refine their methods in response to setbacks. The study emphasizes that understanding these roadblocks and attacker reactions provides valuable insights for improving cybersecurity defenses.
A sophisticated phishing campaign impersonating the Indian Income Tax Department has been targeting local businesses. The attack begins with a spear-phishing email containing a PDF attachment that directs victims to a fake compliance portal. This triggers the download of a malicious ZIP file, which initiates a multi-stage infection chain. The payload, delivered through NSIS installers, deploys a Remote Access Trojan (RAT) with persistence capabilities. The malware harvests system information and establishes communication with command and control servers. Technical indicators suggest a China-linked development environment. This campaign demonstrates how seemingly simple tax-themed phishing can lead to complete device compromise, emphasizing the need for heightened security awareness.
The 'Artemis' campaign, conducted by APT37, utilizes malicious HWP documents with embedded OLE objects to initiate attacks. The threat actor impersonates legitimate entities to gain trust before delivering the payload. The attack chain combines HWP execution with DLL side-loading techniques to evade detection. Steganography is employed to conceal malicious code, and legitimate processes are abused to load malicious DLLs. The campaign targets South Korean organizations, exploiting the widespread use of the HWP format. Multiple stages of encryption and decryption are used to obfuscate the final RoKRAT payload. The threat actor leverages cloud services like Yandex and pCloud for command and control infrastructure, complicating detection and attribution efforts.
A sophisticated espionage campaign targeting Indian entities has been identified, masquerading as the Income Tax Department of India. The activity is associated with the SideWinder APT group, which has evolved its toolkit to evade detection by mimicking Chinese enterprise software. The campaign uses DLL side-loading techniques with legitimate Microsoft Defender binaries to bypass EDR, and utilizes public cloud storage and URL shorteners to evade reputation-based detections. The threat actors employ geofencing behavior, focusing on systems in South Asian timezones. The attack chain includes phishing emails, fraudulent websites, and malicious payloads delivered through file-sharing services. The final stage involves a resident agent that beacons to a command-and-control server, mimicking Chinese endpoint tool protocols.
Threat actors are increasingly abusing legitimate Remote Monitoring and Management (RMM) tools such as PDQ and GoTo Resolve to deploy secondary RMM tools like ScreenConnect or SimpleHelp, enabling persistent and stealthy access. Attacks often start with phishing emails or malicious downloads, targeting diverse sectors including real estate, investment firms, and car dealerships. Social engineering lures include holiday-themed messages and fake bid transcripts to trick users into executing malicious payloads. The abuse of multiple RMM tools complicates detection and remediation efforts. Detection and mitigation require a managed Security Operations Center (SOC) with visibility into RMM tool usage and network behavior. The threat is medium severity due to the potential for persistence and lateral movement, but exploitation requires initial user interaction. European organizations using these RMM tools are at risk, especially those in Germany and countries with high adoption of these management platforms. Practical mitigations include strict RMM access controls, multi-factor authentication, monitoring for unusual RMM deployments, and user training against phishing. No known exploits in the wild have been reported yet, but the trend indicates growing attacker interest in RMM abuse for persistence and lateral movement.
A newly discovered China-aligned APT group named LongNosedGoblin has been targeting governmental entities in Southeast Asia and Japan for cyberespionage purposes. The group employs a varied custom toolset consisting mainly of C#/.NET applications and notably uses Group Policy to deploy malware and move laterally across compromised networks. Their main tools include NosyHistorian for collecting browser history, NosyDoor backdoor using cloud services as C&C, and NosyStealer for exfiltrating browser data. The group has been active since at least September 2023 and uses techniques like AMSI bypassing and living-off-the-land tactics. LongNosedGoblin's campaigns involve multiple stages of execution and various malware components, showcasing a sophisticated approach to cyber espionage operations.
Multiple threat clusters, including state-aligned and financially-motivated actors, are utilizing phishing tools to trick users into granting access to Microsoft 365 accounts via OAuth device code authorization. This technique leads to account takeovers, data exfiltration, and further compromises. Threat actors are leveraging the OAuth 2.0 device authorization grant flow to gain unauthorized access by approving various applications. Campaigns often begin with an initial message containing a URL, which initiates an attack sequence using Microsoft's legitimate device authorization process. Tools like SquarePhish2 and Graphish are being used to facilitate these attacks. Both cybercriminal groups and state-aligned actors have adopted this technique, with Russia-aligned threat actors being particularly active.
A new malware distribution campaign utilizing compromised YouTube accounts to spread infostealers has been identified. The campaign employs GachiLoader, a heavily obfuscated Node.js loader, to deploy the Rhadamanthys infostealer. GachiLoader implements anti-analysis techniques and uses a novel PE injection method called Vectored Overloading. To aid analysis, researchers developed an open-source Node.js tracer tool. The campaign has affected over 100 videos with 220,000 views across 39 compromised accounts since December 2024. The malware evades detection, elevates privileges, and disables Windows Defender before retrieving its payload.
A malware campaign discovered on the NuGet package repository targets the cryptocurrency ecosystem by distributing 14 malicious packages impersonating legitimate crypto-related tools. These packages employ techniques such as homoglyphs, version bumping, and inflated download counts to appear trustworthy and evade detection. The malware aims to steal crypto funds by redirecting transactions and exfiltrating secrets, including OAuth tokens for Google Ads accounts. The campaign highlights the risks of software supply chain attacks, especially for projects relying on compromised dependencies. No known exploits in the wild have been reported yet, but the threat poses a significant risk to developers and organizations integrating these packages. The attack affects . NET developers using NuGet packages related to cryptocurrency and OAuth services. The severity is assessed as medium due to the potential confidentiality and financial impact, combined with moderate exploitation complexity. European organizations involved in blockchain development, fintech, and digital advertising are particularly at risk. Mitigation requires strict dependency vetting, use of package integrity verification, and monitoring of OAuth token usage.
Between June 2024 and April 2025, a sustained credential-harvesting campaign targeting UKR.NET users was identified, attributed to the Russian state-sponsored threat group BlueDelta. The group deployed multiple credential-harvesting pages themed as UKR.NET login portals, leveraging free web services and proxy tunneling platforms to collect user credentials. BlueDelta distributed PDF lures with embedded links to evade detection. The campaign demonstrates the group's adaptability and persistent focus on Ukrainian user credentials for intelligence purposes. Infrastructure changes, including the transition to ngrok and Serveo, reflect responses to takedown efforts. The activity highlights the GRU's continued interest in compromising Ukrainian credentials amid ongoing conflict.
A Chinese-nexus advanced persistent threat actor, UAT-9686, is actively targeting Cisco AsyncOS Software for Secure Email Gateway and Secure Email and Web Manager. The campaign, ongoing since late November 2025, exploits non-standard configurations to execute system-level commands and deploy a persistent Python-based backdoor called AquaShell. Additional tools observed include AquaTunnel for reverse SSH tunneling, chisel for TCP/UDP tunneling, and AquaPurge for log clearing. The attackers can execute encoded commands in the system shell and create reverse connections to attacker-controlled servers. This sophisticated attack aligns with tactics used by other Chinese APT groups, raising concerns about potential widespread impact on email security infrastructure.
Parked domains are increasingly being weaponized through direct search advertising, posing significant risks to users. The investigation found that over 90% of visits to parked domains led to scams, malware, or unwanted content. Three key actors were identified: one using lookalike domains and mail collection, another employing sophisticated 'double fast flux' techniques, and a third exploiting DNS configuration typos. These actors actively profile visitors and selectively redirect traffic to malicious advertisers. The complexity of the advertising ecosystem makes it difficult to trace the origin of threats. Recent policy changes and the rise of AI may inadvertently increase risks associated with parked domains.
RansomHouse, a ransomware-as-a-service operated by the Jolly Scorpius group, has upgraded its encryption capabilities with a new version of its Mario ransomware component. This upgrade introduces a sophisticated two-stage encryption process, enhanced memory management, and dynamic file processing, making the ransomware more efficient and harder to analyze. The attack chain involves MrAgent managing deployments and Mario performing file encryption, primarily targeting virtualized environments such as ESXi servers. Although no known exploits are currently in the wild, the improvements signal a trend toward more resilient ransomware variants. European organizations using virtualized infrastructure are at risk, especially those with ESXi deployments. The threat requires no known CVE but poses a medium severity risk due to its complexity and potential impact on availability and data confidentiality. Mitigation should focus on securing virtualization platforms, monitoring for indicators of compromise, and implementing robust backup and recovery strategies. Countries with high virtualization adoption and critical infrastructure reliance on virtual environments are most likely to be affected.
The ForumTroll APT group has launched a new targeted phishing campaign against Russian political scientists, exploiting plagiarism reports as bait. The attackers used sophisticated techniques, including a well-prepared domain and personalized emails, to deliver the Tuoni framework malware. This campaign follows their spring attacks, which targeted organizations using zero-day vulnerabilities. The fall campaign relied on social engineering, using emails posing as a scientific library to trick victims into downloading malicious archives. The final payload was delivered through a PowerShell script and established persistence using COM Hijacking. Despite being less technically sophisticated than the spring campaign, this operation demonstrates the group's continued focus on Russian and Belarusian targets.
A spear phishing campaign targeting a Colombian government agency under the Ministry of Commerce, Industry and Tourism was discovered in September 2025. The attack, attributed to BlindEagle, utilized a compromised email account within the organization to bypass security controls. The campaign employed a sophisticated multi-layer attack chain, including a fake web portal, nested JavaScript and PowerShell scripts, steganography, and the deployment of Caminho as a downloader for DCRAT. The attack leveraged legal-themed lures, in-memory execution, and abuse of legitimate services like Discord. BlindEagle's evolution in tactics and use of new tools like Caminho demonstrates their ongoing threat to Colombian institutions.
