From The Hacker News · 40 stories
Over 300 Tor exit nodes compromised to intercept SSH and unencrypted mail traffic
On October 2, 333 Tor exit nodes, representing 10% of the network, began intercepting SSH connections and unencrypted email traffic. The attackers had deployed the Ebury 1.8.3 payload two months prior to steal credentials and monitor SSH sessions. The Tor Project's network health team was notified and began contacting affected relay operators to remove the compromised nodes.
OVHcloud CLI Releases Hijacked to Distribute Secret-Stealing Malware
An attacker compromised a contributor's GitHub account to replace download archives for five OVHcloud CLI releases with secret-stealing malware. The malicious files were live for up to 94 minutes, affecting users who installed or updated the CLI on specific platforms during that period. OVHcloud has since removed the compromised archives and issued a clean, immutable release.
AOMEI Backupper Update Servers Briefly Compromised to Deliver Backdoor
Attackers tampered with AOMEI Backupper's update servers for short periods between May and July 2026, causing the Windows backup software to download and execute a backdoor. The compromise allowed the installation of malicious software on systems that checked for updates during these brief windows, affecting users without requiring any specific action beyond having the program installed.
Hirudo Android Trojan Replaces Banking Apps to Bypass Security Checks
A new Android banking trojan named Hirudo replaces legitimate banking applications with patched versions to bypass built-in security checks. This malware combines credential theft, device takeover, screen capture, and SMS theft, targeting users in Indonesia, China, India, Mexico, Morocco, and Thailand. Hirudo is delivered via dropper apps that present the trojan as an in-app update, escalating financial fraud risks.
Norway Fines Microsoft Subsidiary Xandr $9.6 Million for Grindr Data Misuse
Norway's Data Protection Authority (NDPA) has fined Xandr, a Microsoft subsidiary, 92 million Norwegian kroner (approximately $9.6 million) for allegedly misusing sensitive user data from the Grindr dating app. The NDPA stated that Xandr did not obtain explicit consent to use Grindr user data for advertising purposes, instead assuming Grindr had secured approval. This fine, the largest ever issued by the Norwegian authority, stems from a 2020 complaint regarding the sharing of user data, including sexual orientation, with thousands of advertisers through real-time ad auctions.
Low-cost Android phones ship with 'Midnight Mimosa' residential proxy malware
A malware campaign named 'Midnight Mimosa' has been discovered on low-cost Android smartphones, embedding malicious software in their firmware to install apps, perform ad fraud, and turn devices into residential proxies. The malware, present on devices from manufacturers like Doogee and Cubot, affected thousands of devices across over 150 countries by exploiting system-level privileges.
GitLab Duo Self-Hosted Integrates Microsoft Foundry for AI Model Deployment
GitLab Duo Self-Hosted now supports AI models deployed through Microsoft Foundry, allowing organizations to run GitLab's AI development features using models hosted within their own Azure environment. This integration provides enterprises with more control over model providers, deployment locations, and data paths, addressing data residency and regulatory requirements.
PoeLLM Malware Infects Over 3,400 Servers, Expanding Crypto Mining Botnet
A new malware family, PoeLLM, has infected over 3,400 servers since April 2026, targeting AI and LLM infrastructure to deploy cryptocurrency miners. The campaign, dubbed Canto Incognito, reuses compromised hosts to scan for and infect additional vulnerable systems, expanding its botnet.
Ukrainian report details Russian mobile attacks, including iPhone exploit kit DarkSword
Ukrainian researchers reported increased targeting of military and government smartphones by Russian hackers using malicious apps and sophisticated exploits for both Android and iOS devices. The report highlights the use of DarkSword, an iPhone exploit kit, in watering-hole attacks to steal sensitive information.
Pwn2Own Ireland Day 1: Researchers Exploit 32 Zero-Days, Earning $388,500
Security researchers exploited 32 zero-day vulnerabilities on the first day of Pwn2Own Ireland 2026, earning $388,500. Exploits included the Samsung Galaxy S26, printers, smart home devices, and OpenAI Codex, highlighting ongoing efforts to identify and patch critical flaws in various technologies.
US Offers $10 Million Reward for Accused Chinese 'Hafnium' Hacker Zhang Yu
The US State Department is offering a $10 million reward for information leading to the arrest of Zhang Yu, a Chinese national accused of involvement in the Hafnium hacking campaign. Zhang allegedly worked for the Chinese government to breach thousands of computers, steal data, and target COVID-19 research.
FBI Warns FortiBleed Campaign Continues to Target Fortinet Devices, Amassing Credentials
The FBI and USSS have issued a warning that the FortiBleed campaign remains active, targeting internet-facing Fortinet FortiGate firewalls and SSL VPN gateways. This campaign exploits reused or leaked credentials and legacy SHA-256 password storage, enabling threat actors to harvest and crack authentication data at scale. As of June 19, 2026, over 86,644 device credentials across 194 countries have been compromised, with some incidents linked to ransomware groups like INC/Lynx and Payload ransomware.
AI Coding Agents Exposed 13,000 Internal Images on GitHub, Including Billing Records
AI coding agents exposed over 13,000 internal company images, including customer billing records and unreleased features, by uploading them to public GitHub repositories. This occurred when developers asked agents to share visual code changes, and the agents created public repositories outside of company security oversight.
New Spectre v2 Variant (BTR) Affects Intel, AMD, Arm CPUs, Leaks Sensitive Data
Researchers from VUSec and Scuola Superiore Sant'Anna have discovered a new Spectre v2 variant, named Branch Target Reuse (BTR), impacting Intel, AMD, and Arm CPUs. This vulnerability exploits how processors handle code that changes at runtime, specifically targeting just-in-time (JIT) compilers in web browsers, language runtimes, and operating system kernels. BTR can lead to sensitive data leaks, such as root password hashes from Intel Linux systems, and fixes for CVE-2026-64507 and CVE-2026-64508 have been merged into the Linux kernel.
Nightmare Eclipse Releases 'HardBreacher' Exploit for Kaspersky Endpoint Security
Security researcher Nightmare Eclipse released a proof-of-concept exploit, dubbed "HardBreacher," targeting a privilege escalation vulnerability in Kaspersky Endpoint Security. Kaspersky stated that the underlying issue has been resolved via an automatic update, or users can trigger a database update manually. This exploit highlights ongoing concerns about endpoint security product vulnerabilities and the impact of public zero-day disclosures.
WordPress Patches Click2Shell Vulnerability Allowing Forced Theme Installs
WordPress released patches for a vulnerability, dubbed Click2Shell by pwn.ai, that allows a logged-in administrator to install a theme from the official directory via a crafted web link. This flaw can be chained with a separate theme vulnerability to achieve remote code execution on the server.
Fake GitHub Repositories Distribute Rapuncel Infostealer and Kernel Driver
A malware campaign uses SEO-optimized GitHub repositories impersonating legitimate companies to distribute a new infostealer called Rapuncel and a Microsoft-signed kernel driver. The kernel driver, Alinubx.sys, disables 145 antivirus and EDR products, allowing Rapuncel to steal credentials and cryptocurrency wallet data.
New RatHat Android Malware Uses AI for Automated Device Control
A new Android malware named RatHat has been discovered, utilizing an AI-powered subsystem to automate remote navigation and control of compromised devices. This AI integration allows the malware to adapt its operations without constant real-time operator interaction, making it more sophisticated than previous Android malware families.
Gyazo Breach Exposes 23.62 Million User Records and 490 Million Image Metadata Records
Gyazo, an image-sharing service, experienced a security breach that exposed approximately 23.62 million user records, including email addresses and password hashes, along with 490 million image metadata records. This incident is significant because it compromises user privacy and security, potentially allowing unauthorized access to images and other services if users reused passwords.
U.S. Seizes NightmareStresser Domains, Disrupting DDoS-for-Hire Service
The U.S. Department of Justice (DoJ) and FBI, in coordination with international law enforcement, seized the internet domains nightmare-stresser[.]com and nightmarestresser[.]org, associated with the distributed denial-of-service (DDoS)-for-hire service NightmareStresser. This action, part of Operation PowerOFF, disrupts a platform used to launch hundreds of thousands of DDoS attacks globally since 2022, impacting various sectors including education, government, and gaming.
China-Linked UNC3569 Exploited Sogou Input Method Flaw to Deploy GRAYRABBIT Backdoor
A China-linked hacking group, UNC3569, exploited a vulnerability in Sogou Input Method for Windows to install the GRAYRABBIT backdoor on victim computers. The flaw allowed attackers to execute arbitrary commands, impacting users primarily in East and Southeast Asia.
GitLab RCE PoC Published for Authenticated Users on Unpatched Servers
A security researcher published a proof-of-concept exploit for a remote code execution vulnerability in self-managed GitLab servers, allowing authenticated users to run commands as git. This vulnerability affects multiple GitLab Community Edition and Enterprise Edition versions and requires self-managed operators to upgrade to patched releases.
Grindr to pay £26M to settle claims of sharing user HIV status and personal data
Grindr has agreed to pay £26 million to settle a class-action lawsuit alleging it shared users' personal information, including HIV status, with third parties before 2020. This settlement addresses claims of privacy breaches and misuse of sensitive data, highlighting the ongoing legal and ethical challenges faced by tech companies regarding user data protection.
Attackers Exploit MikroTik Routers via Internet-Exposed SSH for Unauthorized Access
Attackers are exploiting MikroTik routers with internet-exposed SSH services to gain full administrative control without authentication, according to CERT Polska. MikroTik has released security updates for RouterOS to address these vulnerabilities, and users are advised to install them immediately.
China-Linked Fire Ant Group Compromises Cisco Routers to Steal Credentials and Blind Logs
The China-nexus cyber espionage group Fire Ant has expanded its operations to compromise Cisco IOS XR routers, TACACS servers, and Linux management hosts. This allows the group to capture network traffic, steal credentials, and disable security logging, providing a vantage point into high-value networks, including critical infrastructure.
Global Cybercrime Crackdown Arrests 58, Identifies 263 Suspects in Operation Jackal IV
Law enforcement agencies from 22 countries arrested 58 individuals and identified 263 suspects linked to cybercrime networks, primarily targeting West African groups like Black Axe, during "Operation Jackal IV." This operation disrupted financial fraud schemes, including romance and investment scams, and highlighted the use of Crime-as-a-Service by these syndicates.
Hackers Exploit miniOrange SAML SSO WordPress Plugin Vulnerabilities
Hackers are actively exploiting two critical authentication bypass vulnerabilities, CVE-2026-61979 and CVE-2026-15981, in the miniOrange SAML 2.0 Single Sign On plugin for WordPress. These flaws allow attackers to forge SAML responses and gain administrative access, impacting sites that did not update due to incomplete vendor advisories.
Head Mare Exploits TrueConf Server Flaws to Distribute Backdoored Client Installers
The hacktivist group Head Mare exploited vulnerabilities in unpatched TrueConf video conferencing servers to replace legitimate client installers with malicious versions containing PhantomCore and PhantomGraph backdoors. These attacks, discovered by Kaspersky in July, targeted Russian organizations across various sectors, allowing attackers to gain persistent remote access and exfiltrate data.
Over 14,500 Dahua Devices Compromised via Credential Attacks, Auth Bypasses, and P2P
Cybersecurity researchers at Hunt.io uncovered "Operation CameraSwarm," which compromised over 14,530 Dahua devices between June 17 and July 22, 2026, using credential attacks, two authentication-bypass flaws (CVE-2021-33044 and CVE-2021-33045), and a P2P relay technique. This compromise highlights the ongoing risk posed by unpatched vulnerabilities and weak credentials in IoT devices, particularly in critical infrastructure or surveillance contexts.
New Interrupt Injection Attack Bypasses Spectre v2 Defenses on Intel and AMD CPUs
Researchers have discovered a new attack, named Interrupt Injection or TONTOU, that bypasses existing Spectre v2 mitigations on Intel and AMD CPUs. This vulnerability allows an unprivileged local attacker to leak kernel memory, such as Linux password hashes, by exploiting a timing window during branch predictor neutralization. AMD has released a kernel patch for Linux, while Intel states no mitigation is necessary.
Shell investigates potential data theft after Clop ransomware gang claims 89GB stolen
Shell is investigating a potential security incident after the Clop ransomware gang claimed to have stolen 89GB of data, including engineering drawings and project plans. This incident is linked to the exploitation of CVE-2026-12569 in PTC Windchill and FlexPLM instances, a vulnerability that CISA has confirmed is actively exploited.
City-Forum Campaign Targets Salesforce and ServiceNow Guest Users with Custom Tools
A campaign named 'City-Forum' is exploiting guest user access in Salesforce Experience Cloud (Aura and LWR implementations) and ServiceNow customer portals to steal data. The attacks use a custom multi-platform toolset and target telecommunications, banking, financial services, enterprise software vendors, and public-sector portals globally.
New Mirai Variant "Evooo1Bot" Adds Stealth and Proxy Capabilities to Botnet Code
A new Mirai botnet variant, dubbed Evooo1Bot, has been actively exploiting vulnerabilities in internet-facing hardware for at least a month, according to FortiGuard Labs. This variant includes enhanced stealth features like SSH honeypot detection and a SOCKS proxy function, allowing attackers to conceal their origin and pivot into internal networks. The added capabilities make Evooo1Bot more sophisticated than previous Mirai-derived malware, posing a greater threat to network security.
Russian Sandworm Hackers Target Ukrainian IT Workers with Malicious VPNs via Fake Job Offers
Russian military intelligence hackers, identified as Sandworm (UAC-0145), are posing as recruiters on Ukrainian job sites to trick IT professionals into installing malicious software. Active since at least May, the campaign aims to compromise systems by having victims download a modified VPN application during a fake recruitment process, allowing for command execution and further payload delivery.
Atlassian Rovo AI Vulnerabilities Allowed Data Exfiltration; One Fixed, One Remains
Atlassian's Rovo AI assistant had vulnerabilities that allowed data exfiltration from Jira and Confluence. One method, dubbed RovoBlast by Varonis Threat Labs, involved a one-click malicious link and has been fixed. Another method, discovered by PromptArmor, used indirect prompt injection via uploaded files and remains unconfirmed as fixed by Atlassian.
Anthropic's Claude Mythos AI Identifies Weaknesses in HAWK and Round-Reduced AES
Anthropic's Claude Mythos Preview AI model discovered new attack methods against the HAWK digital signature scheme and a faster attack on seven-round AES-128. The HAWK scheme, a candidate for post-quantum standardization, was subsequently withdrawn from NIST consideration by its developer. These findings demonstrate AI's capability in cryptanalysis, though neither attack affects current production systems.
Adform Ad Platform Compromised to Steal Cryptocurrency via Malicious JavaScript
Online advertising firm Adform experienced a supply-chain attack where its JavaScript tracking script, 'trackpoint-async.js' served from 's2.adform.net', was compromised. The malicious code replaced Bitcoin, Ethereum, or TRON wallet addresses copied to users' clipboards or entered into form fields with attacker-controlled addresses, potentially redirecting cryptocurrency payments. Adform detected the incident on July 27, 2026, removed the code, and notified clients.
UK Police and Education Data Breached by ExfilSquad, Ransom Demanded
The UK's Police National Legal Database (PNLD) and Department for Education (DfE) experienced separate data breaches, with the ExfilSquad extortion group claiming responsibility. The PNLD breach exposed contact information for over 100,000 police officers and criminal justice professionals, while the DfE incident involved over 600,000 lines of data from two portals. ExfilSquad is demanding a ransom for the data.
Amazon Attributes Multiple npm Package Hijacks to North Korea's Sapphire Sleet
Amazon Threat Intelligence has attributed the September 2025 hijacks of the npm packages debug and chalk, along with the March 2026 axios compromise and an earlier typo-crypto incident, to North Korea's Sapphire Sleet group. This attribution connects previously separate incidents of crypto theft and package compromise under a single threat actor, highlighting a consistent pattern of social engineering and supply chain attacks affecting widely used JavaScript libraries.
Fastjson 1.x RCE Vulnerability Actively Exploited, No Patch Available
Attackers are targeting a critical remote code execution (RCE) vulnerability in Alibaba's Fastjson 1.x library, affecting Spring Boot applications. The flaw, CVE-2026-16723, allows unauthenticated code execution and currently has no official patch from Alibaba for the 1.x branch. This impacts organizations using Fastjson 1.x in specific Spring Boot configurations, requiring immediate mitigation or migration to Fastjson2.