In August 2026, a new variant of the CoffeeLoader malware was identified, distributed via the Retadup USB worm. This variant uses an NSIS-based installer that drops a patched 7z.dll containing malicious shellcode and employs the updated Armoury packer with OpenGL for GPU-based decryption. It establishes persistence through registry modifications and downloads its main payload from hardcoded command-and-control servers over HTTPS. The Retadup worm, originally discovered in 2017 and neutralized in 2019, has been resurrected by threat actors who re-registered its former C2 domains to distribute malware, primarily affecting systems in Peru, Angola, Russia, and Bolivia.
GhostFrame is a post-exploitation framework targeting VMware ESXi hypervisors, enabling attackers with root access to control multiple virtual machines, harvest credentials, execute commands, and pivot within victim networks. It has been linked to the TAG-218 threat actor group, whose operations overlap with Settra ransomware. The implant establishes persistence, disables security controls, and uses distinctive command and control communications. It integrates credential harvesting tools such as VMkatz and has been observed following compromises of VPN appliances in the United States and Europe.
LiteLLM version 1.107.0 was compromised and distributed via PyPI and official container images on October 21, 2026. The malicious package used Python startup hooks to automatically execute credential-stealing payloads without requiring explicit import of LiteLLM. Attackers stole developer GitHub credentials to redirect release workflows to a fake repository. The payload targeted authentication credentials across developer tools, AI applications, cloud environments, browsers, and infrastructure. Credentials were encrypted and exfiltrated to attacker-controlled servers. The compromise affected developer workstations, CI/CD agents, test systems, containers, and LiteLLM gateway deployments where Python was executed.
This analysis covers advanced Android threats that operate beyond the standard application sandbox, leveraging new OS-level telemetry features. Android's Delegated Scope allows enterprises to grant security apps access to logs on managed devices without full MDM control. The Intrusion Logging feature extends forensic visibility to BYOD users with Advanced Protection on Android 16 or later. This capability enables collection and sharing of encrypted security, DNS, and connection logs for threat analysis. The analysis reveals the full RatHat attack chain, including Wireless Debugging self-pairing, ADB shell command execution, daemon staging, and command-and-control communications, providing actionable indicators of compromise.
An autonomous offensive agent named 'Skynet' hijacks exposed large language model (LLM) endpoints on the internet to conduct a range of attacks. It uses 74 offensive tools targeting web frameworks, credential theft, database access, and cryptocurrency mining. The agent operates unattended for extended periods by bypassing AI safety controls through jailbreak techniques. Victims include AI model owners who incur costs for computing power and compromised servers exploited for Monero mining. This activity was observed via an AI honeypot capturing the full attack sequence, highlighting a novel self-directed cybercrime targeting unauthenticated internet-facing AI services.
CVE-2026-102764 is a critical authentication bypass vulnerability in BeyondTrust Remote Support versions 26.2.3 and earlier. It enables unauthenticated remote attackers to perform arbitrary file operations without user interaction. Exploitation in the wild has been confirmed, involving deployment of web shells and tunneling tools, root privilege escalation, credential vault extraction attempts, persistence mechanisms, and log deletion. The threat actor ARMORY CASTLE, linked to PRC-nexus espionage, is attributed to these attacks. The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has added this vulnerability to its Known Exploited Vulnerabilities list. Emergency patching and thorough compromise investigations are required, as patching alone does not remove attacker persistence.
CARBONATO is a Docker botnet discovered by ThreatDown researchers that targets exposed Docker daemons on port 2375. It operates by distributing trojanized cryptocurrency wallet applications and running a botnet infrastructure. The botnet uses a modified Hermes Agent, an open-source AI framework, to execute commands via Telegram, maintain persistence, and harvest credentials. The campaign infrastructure spans multiple hosting providers and has been active from October 2024 through August 2026. No CVE identifiers or official patches are associated with this threat. The severity is medium due to credential theft and botnet control capabilities.
The PolinRider campaign compromises developer GitHub accounts to inject malicious code into development versions of popular packages, including visanduma/nova-two-factor on Packagist. It targets multiple ecosystems such as npm, PyPI, Go modules, Packagist, and Chrome extensions by abusing Git-based collaboration workflows rather than directly compromising package registries. The attackers use advanced techniques like rewriting Git history, hiding payloads in configuration and font files, and triggering execution via IDE tasks. PHP projects are specifically targeted with obfuscated JavaScript executed through shell_exec calls. The campaign is attributed to North Korean actors focusing on cryptocurrency theft.
Mythic is an open-source command-and-control (C2) framework used by red teams and threat actors. Analysis found 131 unique Mythic C2 hosts exposed on the public Internet, with most using default TLS certificates. These deployments span major cloud providers and are concentrated in the US, Hong Kong, and China. Default artifacts such as TLS certificates with O=Mythic and port 7443 responses enable detection. Some clusters appear to be training environments, while isolated deployments with custom domains and advanced transport methods like Discord and steganography indicate operational use.
Two sophisticated phishing campaigns employed browser-in-the-browser (BiTB) techniques to deceive victims into installing rogue ScreenConnect remote management tools. Attackers sent phishing messages with malicious links redirecting targets to fake Adobe Reader update pages. The BiTB technique created convincing fake browser windows within webpages, displaying legitimate-looking Adobe URLs to bypass user awareness training. Victims were tricked into downloading ScreenConnect installers disguised as Adobe software updates. Each incident resulted in deployment of multiple rogue ScreenConnect instances for redundant persistence, followed by execution of defense-evasion binaries (HideCursor.exe and HideUL.exe) designed to hide attacker activities. The attacks established service-based persistence through Windows services, enabling continued remote access. Both campaigns were intercepted before further damage occurred, demonstrating how threat actors combine social engineering throughout the entire attack chain...
A previously undocumented Linux toolkit has been targeting South Korean automotive and media organizations with minimal detection since early 2025. The campaign employs a HAProxy instance called ted backdoor, compiled within the victim's existing HAProxy version 2.8.12, alongside trojanized versions of crond, agetty, atd, sshd, and polkitd. This sophisticated framework enables remote command execution, malicious script injection into web traffic, credential harvesting, and long-term surveillance. The ted backdoor uses HAProxy's native filter API and internal structures to intercept SSL-decrypted HTTP traffic while maintaining legitimate load balancing operations. Operating alongside are an SSH keylogger, a curl-based RAT with HAProxy health monitoring capabilities, and a deployment stager. The toolkit is attributed with medium confidence to DPRK APTs based on targeting patterns, simple XOR-based encryption schemes, custom substitution ciphers, and C2 infrastructure associated with APT37.
Beginning in 2024, a financially motivated threat actor designated BREEZE COMET has conducted sophisticated operations targeting Brazilian financial services, retail, and eCommerce organizations. The group specializes in manipulating payment systems including Pix, STR, and Boleto to conduct fraudulent transfers worth tens of thousands of USD. Their evolved tactics leverage customized malware suites written in multiple languages including Rust, Nim, Golang, and Java, alongside compromised government websites for initial access and command and control. The threat actor demonstrates advanced capabilities by targeting banking software, payment APIs, and mTLS credentials while maintaining persistent access through multiple backdoors. Evidence indicates BREEZE COMET uses generative AI to accelerate malware development and script creation, suggesting potential expansion to other Latin American and African countries based on infrastructure replication observed in Nigeria, Paraguay, Ghana, and Venezuela.
In July 2026, Microsoft Threat Intelligence uncovered a supply chain attack targeting the official AsyncAPI NPM organization. Attackers published malicious versions of multiple packages under the trusted AsyncAPI namespace, exploiting developer dependencies to distribute malware. The compromised packages deployed a multi-stage Remote Access Trojan through obfuscated lifecycle hooks that executed during routine build workflows. Upon installation, the malware retrieved second-stage payloads from IPFS gateways, established persistence on infected systems, and initiated command-and-control communications with external infrastructure. The attack leveraged trusted build automation and dynamic package retrieval via npx to bypass traditional security controls, affecting developers executing version-pinned tasks in their CI/CD pipelines.
This analysis discusses how poorly designed AI guardrails in security operations can unintentionally aid attackers by impeding defensive actions. Overly restrictive or inflexible AI filters controlled by third-party providers may cause delays or refusals in security investigations, giving adversaries more time to complete their objectives. The author recommends that security teams maintain operational sovereignty over guardrails, allowing customization and temporary adjustments to safeguards to better align with specific threat models. This flexibility is essential to prevent attackers from exploiting rigid controls to disrupt incident response processes.
The Gentlemen is a ransomware group active since July 2025, operating a Ransomware-as-a-Service model with dual-extortion tactics. They target Windows, Linux, and ESXi systems, focusing on extensive preparation before encrypting data. Their methods include privilege escalation using legitimate tools, persistence via registry and scheduled tasks, disabling security tools, deleting logs, and terminating backup services. They use strong encryption algorithms XChaCha20 and Curve25519. The group primarily targets medium-to-large organizations in the Asia-Pacific region, with a recent surge in activity. Victims face ransom demands with about 10-day deadlines and threats of data publication if unpaid.
In May 2026, a cyber-espionage campaign by the Armored Likho group (also known as Eagle Werewolf) targeted private individuals and organizations across Russia, including major corporations, public sector entities, IT companies, and educational institutions. The attackers employed fake donation service applications as initial infection vectors. The campaign introduced the Still Toolkit, comprising two Rust-based components: Still Sync, which steals Telegram session data and leverages the Telegram API to extract chat logs and media files, and Still Audio, an implant that conducts covert audio surveillance by detecting speech patterns and recording conversations. The toolkit demonstrates sophisticated capabilities including Dead Drop Resolver techniques, RMS-based voice activity detection, and gRPC-based C2 communications. The campaign shows significant code overlap with previous Armored Likho operations, particularly from February 2026, including identical dropper architecture, encryption algorithms, and inf...
Threat actors are conducting a multi-wave campaign using social engineering lures themed around Zoom updates, business documents, and system utilities to deploy ScreenConnect Remote Monitoring and Management agents. The operation employs VBScript droppers, batch loaders, compiled .NET executables, and HTML phishing pages, all retrieving payloads from a WsgiDAV staging server at 207.174.0.143:8080. Victims receive silently installed ScreenConnect agents that beacon to three attacker-controlled relay servers, providing persistent remote access. The campaign demonstrates technical evolution from obfuscated VBScript with XOR encryption to aggressive .NET loaders executing nine-step Windows Defender destruction sequences. Cross-platform variants target both Windows and macOS systems. All payloads are legitimately signed ConnectWise ScreenConnect MSIs, designed to evade security controls that trust code signing. The threat actor actively rotates payload hashes and recently pivoted to stealth tactics specifically...
A sophisticated supply chain attack compromised the legitimate 3CXDesktopApp softphone application across Windows, macOS, and Linux platforms. The malicious activity involved trojanized signed installers that deployed a compromised ffmpeg.dll binary, establishing HTTPS beacons to attacker-controlled infrastructure and enabling second-stage payload deployment. Analysis revealed the attack utilized specific beacon structures and encryption keys matching infrastructure patterns, with hands-on-keyboard activity observed in targeted cases. The operation affected multiple platforms through signed MSI installers containing malicious components. The attack demonstrated advanced tradecraft through abuse of trusted software distribution channels, requiring immediate removal of affected versions and deployment of behavioral detection capabilities to identify malicious beaconing activity.
Since late 2025, cybercriminals have been exploiting Wallpaper Engine, a popular live wallpaper application on Steam, to distribute malware through Steam Workshop. Attackers target primarily Chinese and Russian gamers by embedding malicious code within application wallpapers shared on the platform. These compromised wallpapers deliver various malware types including infostealers, backdoors, crypto miners, and ransomware. One analyzed sample dropped DarkKomet backdoor while hijacking Steam sessions to steal account credentials. The malware modifies system libraries to locate Steam installations and exfiltrate data to attacker-controlled servers. Compromised accounts are then used to upload additional malicious wallpapers. The diverse malware families suggest multiple independent hacking groups are exploiting this distribution method. Infected wallpapers received thousands of downloads before removal, with 89% of infections occurring in China.
A long-running typosquatting campaign impersonated the widely used shopspring/decimal Go library by publishing github.com/shopsprint/decimal, differing by a single character. Active since November 2017, the package remained benign through seven releases until being weaponized in August 2023 with version v1.3.3. This version introduced a malicious init() function that executes automatically on import, establishing a DNS TXT record-based command and control channel to dnslog-cdn-images.freemyip.com. The backdoor polls every five minutes and executes arbitrary commands returned via TXT records. Although the GitHub repository and owner account have been deleted, the malicious module remains permanently cached and accessible through Go's module proxy system, continuing to pose a supply chain risk to developers who mistype the package name.