Intelligence Feed

3mo ago · bleeping-computer

A critical 16-year-old Linux kernel vulnerability named Januscape (CVE-2026-53359) enables guest-to-host virtual machine escape on both Intel and AMD architectures, posing significant risks to multi-tenant cloud environments. The flaw resides in KVM/x86's shadow MMU emulation and can be exploited by attackers with root access inside a guest VM to execute arbitrary code on the host or cause a denial-of-service by crashing the host kernel. It was used as a zero-day in Google's kvmCTF program, and when combined with the Dirty Frag privilege escalation, even unprivileged attackers could achieve full compromise. Administrators are urged to apply patch commit 81ccda30b4e8 to mitigate exposure.

3mo ago · bleeping-computer

Modern email attacks increasingly exploit trusted identities, legitimate services, and normal communication patterns to bypass traditional security controls. Techniques such as Device Code phishing, business email compromise (BEC), and account takeovers (ATO) are commonly used, allowing attackers to evade detection by avoiding malware and suspicious attachments. These attacks result in alert fatigue and manual investigation overhead for security teams. The article highlights the need for behavioral AI to automate detection and response in the evolving email threat landscape.

3mo ago · hacker-news

BeyondTrust has patched multiple critical vulnerabilities in its Remote Support (RS) and Privileged Remote Access (PRA) products. The most severe flaws, CVE-2026-40138 and CVE-2026-40139, allow unauthenticated attackers to bypass authentication and gain unauthorized access to appliances, including elevated privilege accounts, under specific configurations. Additional vulnerabilities could lead to denial-of-service conditions and unauthorized data access. While no active exploitation has been reported, past flaws in these products have been actively exploited to deploy web shells and backdoors, underscoring the urgency of patching.

3mo ago · hacker-news

CERT/CC has warned of a hidden administrative backdoor in multiple Tenda router firmware versions, tracked as CVE-2026-11405. The backdoor resides in the '/bin/httpd' binary and allows full administrative access by bypassing normal authentication using an undocumented 'rzadmin' password check. The vulnerability enables attackers to gain elevated privileges without valid credentials, leading to potential device takeover. No patch is currently available, and users are advised to disable remote management and change default LAN IP settings.

3mo ago · bleeping-computer

Microsoft is testing a new Cloud Rebuild feature for Windows 11, designed to remotely reinstall the operating system on devices that are non-bootable or experiencing critical issues. This feature, part of the Windows Resiliency Initiative, allows for a full OS recovery without the need for physical media or a functioning OS. It is being rolled out to Windows Insider Experimental channel users as part of ongoing efforts to improve system recovery and resilience.

3mo ago · bleeping-computer

BeyondTrust has disclosed two critical vulnerabilities, CVE-2026-40138 and CVE-2026-40139, in its Remote Support (RS) and Privileged Remote Access (PRA) software that allow unauthenticated attackers to bypass authentication and gain unauthorized access to affected systems. Exploitation requires specific configurations, though details were not disclosed. The vendor has patched cloud instances and urges self-hosted customers to update to version 25.3.3 or later. Previous vulnerabilities in the same software have been actively exploited in the wild by threat actors, including the Chinese state-backed group Silk Typhoon.

3mo ago · bleeping-computer

Microsoft is enabling the Windows settings backup and restore tool by default for enterprise devices joining or hybrid-joining Microsoft Entra, starting with Windows 11 version 26H2. This feature, previously opt-in, aims to streamline user settings backup during device resets, replacements, or upgrades. The default enablement applies only to eligible non-EU DMA-regulated devices where administrators have not explicitly configured backup policies. IT admins retain control via MDM solutions like Microsoft Intune or Group Policy to override defaults.

3mo ago · hacker-news

A critical 16-year-old use-after-free vulnerability in Linux's KVM hypervisor, dubbed 'Januscape' and tracked as CVE-2026-53359, allows guest virtual machines to escape to the host on both Intel and AMD x86 systems. The flaw resides in the shadow MMU code and can lead to host kernel corruption or full host code execution when exploited. The vulnerability affects any x86 environment hosting untrusted guests with nested virtualization enabled, posing a high risk to multi-tenant cloud environments.

3mo ago · hacker-news

An Iranian hacking group linked to the Ministry of Intelligence and Security (MOIS), tracked as Cavern Manticore, has been using a new modular command-and-control (C2) framework named Cavern to target Israeli organizations, particularly in the IT and government sectors. The attack leverages DLL side-loading via SysAid's software update mechanism, deploying a trojanized DLL (uxtheme.dll) that communicates with a C2 server and downloads additional malicious modules. These modules enable reconnaissance, data theft, lateral movement, and tunneling, with a sophisticated .NET-based architecture using mixed compilation formats to hinder analysis. The group exploits trusted relationships in the software supply chain and has shifted from broad reconnaissance to targeted data exfiltration across Middle Eastern sectors.

3mo ago · bleeping-computer

Vietnamese authorities have arrested seven suspects linked to the operation of HiAnime, a major anime piracy service that attracted hundreds of millions of visitors monthly. The group allegedly created over 100 websites to distribute more than 26,000 pirated anime titles, generating $12.85 million in illicit advertising revenue between 2020 and April 2026. The service was previously rebranded from Zoro.to to Aniwatch.to and later to HiAnime.to, appearing on both the European Commission's and USTR's notorious markets lists. The Alliance for Creativity and Entertainment (ACE) commended the takedown, which resulted from a multi-year investigation supported by U.S. law enforcement.

3mo ago · bleeping-computer

Threat actors are conducting fake IT support calls via Microsoft Teams to trick employees into installing the EtherRAT malware. The attack begins with a phishing email containing a malicious PDF, followed by a voice call from an external Teams account impersonating system administrators. Attackers use legitimate remote access tools like HopToDesk and AnyDesk, then deploy EtherRAT via a malicious MSI installer, enabling full system control and data theft. EtherRAT uses Ethereum smart contracts for C2 resilience, and the campaign is actively evolving with multiple malware versions observed.

3mo ago · bleeping-computer

A phishing campaign impersonates over 30 well-known brands, including Adobe, Netflix, Coca-Cola, and OpenAI, to target marketing professionals with fake job interviews. The attackers abuse legitimate platforms like PeopleForce and Salesforce Marketing Cloud to increase credibility, using nested redirects to deliver a malicious landing page. Victims are prompted to sign into their Google accounts via a fake 'Continue with Google' page that uses browser-in-the-browser (BitB) technique to mimic legitimate authentication, enabling credential theft.

3mo ago · hacker-news

Threat actors are actively probing and attempting to exploit CVE-2026-20896, a critical vulnerability in Gitea Docker images that allows unauthenticated remote users to gain elevated access by spoofing the 'X-WEBAUTH-USER' HTTP header. The flaw arises from the default configuration trusting all IP addresses due to a wildcard in the 'REVERSE_PROXY_TRUSTED_PROXIES' setting. Exploitation attempts were detected just 13 days after public disclosure, with initial activity observed from a ProtonVPN IP address. Although no full exploitation has been confirmed yet, the risk remains high for unpatched internet-facing instances.

3mo ago · hacker-news

A suspected China-nexus threat actor is conducting a targeted cyber espionage campaign against Indian taxpayers, tax professionals, and corporate finance teams using spear-phishing emails impersonating the Indian Income Tax Department. The campaign, dubbed Operation DragonReturn, delivers DcRAT via a malicious fake tax filing utility to steal sensitive data and establish persistent access. The attackers use social engineering, DLL side-loading, image-based payload concealment, and Windows service persistence to maintain long-term access to compromised systems.

3mo ago · hacker-news

The article discusses how to evaluate AI-powered Security Operations Center (SOC) platforms in 2026, emphasizing the importance of real-time data correlation, full-lifecycle automation, and measurable outcomes. It contrasts true AI SOC platforms with 'bolt-on' AI solutions that offer limited automation. The focus is on architectural capabilities rather than marketing terms, with Exaforce's Exabots highlighted as an example of an agentic AI SOC platform performing detection, triage, investigation, and response. No specific malicious threat activity, threat actor, or compromise indicators are described.

3mo ago · hacker-news

The article covers multiple cyber threats including the disruption of the NetNut residential proxy network, which leveraged compromised home devices to route malicious traffic. Threat actors are targeting researchers with fake proof-of-concept repositories delivering the ChocoPoC RAT, while the Scattered Spider group is linked to extortion attempts. New malware such as Ousaban and browser-based ransomware exploit social engineering and legitimate browser APIs, and AI-driven attacks are increasing in sophistication, including indirect prompt injection and fake phishing pages generated via AI.

3mo ago · bleeping-computer

Threat actors are actively exploiting a maximum-severity vulnerability in Adobe ColdFusion, tracked as CVE-2026-48282, to achieve remote code execution on unpatched systems. The flaw affects ColdFusion versions 2025.9, 2023.20, and earlier. Exploitation began within two hours of public disclosure, prompting urgent warnings from KEVIntel and the Canadian Center for Cyber Security. Adobe urges administrators to apply patches immediately to mitigate risk.

3mo ago · bleeping-computer

The article discusses the evolution of software development from Waterfall to Agile, DevOps, and now 'Vibe Coding,' where generative AI enables rapid application creation through natural language prompts. This acceleration introduces significant security risks, as AI-generated code may contain vulnerabilities, architectural flaws, or compliance issues despite appearing functional. Traditional secure development practices like threat modeling, code review, and least privilege remain critical to mitigate these risks. The article warns that without proper governance, Vibe Coding could become a modern form of shadow IT—innovative but potentially dangerous.

3mo ago · socket-dev

The Node.js Security Working Group is considering a shift to a public workflow for handling lower-severity security reports due to an overwhelming influx of AI-generated vulnerability submissions via HackerOne. Security maintainer Rafael Gonzaga attributes the surge to LLM-driven fuzzing and scanning tools, which produce highly similar reports, undermining the value of private disclosure. The proposal aims to reduce operational overhead by reserving private embargo processes for high-severity issues while enabling public handling for others, though maintainers remain divided on whether this change will alleviate workload bottlenecks.

3mo ago · hacker-news

Researchers from the Hong Kong University of Science and Technology have demonstrated SKILLCLOAK, a technique that enables malicious AI agent skills to evade static scanners by using self-extracting packing and obfuscation methods. These cloaked skills hide payloads in ignored directories like .git/ or use character substitution and command splitting to bypass pattern-based detection, remaining undetected in over 90% of cases across eight tested scanners. The study also introduces SKILLDETONATE, a runtime behavioral checker that detects 97% of attacks by monitoring system-level activity, highlighting the insufficiency of static analysis alone. Real-world marketplaces already host malicious skills using similar evasion tactics, indicating active exploitation.

3mo ago · hacker-news

A vulnerability in Opera GX allowed malicious websites to silently install browser mods in the form of .crx files without user interaction, enabling attackers to exfiltrate sensitive data such as Gmail addresses via universal CSS injection. The attack exploited the browser's auto-install feature for mods, which applied malicious CSS rules across all visited sites, facilitating cross-site leak (XS-Leak) techniques. Although Opera patched the flaw in version 130.0.5847.89 and found no evidence of in-the-wild exploitation, the zero-click nature of the attack made it highly effective once triggered.

3mo ago · hacker-news

A new Java-based remote access trojan (RAT) named QuimaRAT has emerged as a malware-as-a-service (MaaS) offering, targeting Windows, Linux, and macOS systems. The malware features a modular architecture with encrypted plugins delivered from command-and-control (C2) infrastructure and supports cross-platform persistence mechanisms. It includes tools for stealthy delivery, such as browser-cache staging and fileless execution, while evading detection through trusted execution paths and obfuscation techniques.

3mo ago · hacker-news

Researchers at Shandong University have demonstrated a new data exfiltration technique called TrojPix, which leverages video cable emissions to leak data from air-gapped systems by modulating on-screen pixels imperceptibly. The method requires prior malware presence on the system but does not need admin rights or hardware modifications, enabling data transmission at up to 8.1 Mbps over distances up to 208 meters. This technique represents a significant advancement in covert channel speed and range, though it remains experimental and not yet observed in real-world attacks.

4mo ago · bleeping-computer

JadePuffer ransomware represents the first documented case of a ransomware operation fully automated by a large language model (LLM) agent. The AI-driven attack exploited CVE-2025-3248 in Langflow to gain initial access, then performed reconnaissance, credential theft, lateral movement, and encryption autonomously. The agent adapted to failures in real time, demonstrating human-like operational resilience and rapid iteration. It encrypted 1,342 Nacos configuration items and left a ransom note with a Proton Mail contact and a Bitcoin address, though the encryption likely used AES-128-ECB rather than AES-256 as claimed.

4mo ago · socket-dev

PolinRider is a North Korea-linked supply chain campaign targeting developer ecosystems, including npm, Packagist, Go modules, and Chrome extensions. The threat actors compromise maintainer accounts, modify legitimate repositories with obfuscated JavaScript loaders, and use Git history rewriting to conceal malicious changes. These loaders retrieve encrypted second-stage payloads from blockchain infrastructure, execute them via eval(), and have delivered malware such as DEV#POPPER and OmniStealer. The campaign remains active, with ongoing compromises across multiple open source platforms.

4mo ago · step-security

A coordinated supply chain attack spanning eight months involved 15 malicious JetBrains plugins that stole AI API keys from approximately 70,000 developers. The plugins, masquerading as legitimate AI coding assistants, exfiltrated OpenAI, DeepSeek, and SiliconFlow API keys to a command-and-control server in Beijing. The stolen credentials were transmitted in plaintext over HTTP, and the attacker's infrastructure remains active despite JetBrains' removal of the plugins and banning of associated accounts.

4mo ago · step-security

On June 24, 2026, a coordinated supply chain attack compromised 20 npm packages within the Leo Platform ecosystem, all published within a 3-second window. The malicious packages contain a CI/CD attack toolkit that steals secrets from GitHub Actions runners, cloud environments, and credential stores, then exfiltrates them using the victim's own GitHub token. The attack uses the same 'Phantom Gyp' technique and obfuscation chain as the earlier Miasma campaign, indicating the same threat actor. These packages collectively receive around 13,600 weekly downloads, posing significant risk to CI/CD pipelines.

4mo ago · socket-dev

GitHub has released actions/checkout v7 to mitigate a long-standing supply chain risk in GitHub Actions where privileged workflows using pull_request_target could execute attacker-controlled code from untrusted pull requests. These workflows run with elevated permissions, including access to secrets and tokens, and previously allowed malicious actors to steal credentials or publish malicious packages. The update blocks unsafe checkouts by default, particularly those pulling code from forked pull requests in high-privilege contexts. This change addresses a known attack pattern exploited in recent incidents involving Nx, PostHog, and TanStack.

4mo ago · socket-dev

The article discusses the growing risk of software supply chain attacks in the era of AI-powered development, where AI agents autonomously pull and execute unvetted open source packages outside traditional security monitoring. These agents operate in blind spots, such as ephemeral sandboxes, where no scanning or registry controls exist. The speed of modern attacks—often exploiting vulnerabilities within hours of disclosure—exceeds traditional response timelines, rendering forensic-focused defenses ineffective. Proactive governance at the point of package ingestion, supported by real-time threat intelligence, is presented as a necessary defense.

4mo ago · socket-dev

The U.S. government abruptly suspended global access to Anthropic's AI models, Claude Fable 5 and Mythos 5, citing national security risks following jailbreaks and unauthorized vulnerability discoveries in federal systems. The models, used for automated code analysis, were deemed a supply chain risk, prompting a federal blackout and export controls under ECRA. The shutdown highlights the growing classification of advanced AI as critical infrastructure, with significant implications for enterprise dependency and national cybersecurity policy.