Find what you’re looking for

ShineyHunters Exploit Zero-Day to Breach Trezor's Shipping Partner

In August 2026, hardware wallet manufacturer Trezor disclosed that 67,000 U.S. customers had their personal data exposed through a breach at shipping provider ShipMonk. The ShineyHunters extortion gang exploited CVE-2026-72898, a critical SQL injection vulnerability in Metabase with a CVSS score of 10.0, to gain unauthorized access to ShipMonk's systems. The exposed data included customer names, email addresses, phone numbers, shipping addresses, and order numbers from November 2019 to August 2021, despite Trezor's repeated requests for data deletion per their 90-day retention policy. This supply chain attack highlights how third-party vulnerabilities can impact customer data even when primary security measures are robust. This incident demonstrates the growing threat of supply chain compromises targeting logistics and fulfillment providers, with attackers increasingly exploiting zero-day vulnerabilities in business intelligence platforms to access customer databases across multiple organizations simultaneously.

Microsoft Warns of Critical Security Gaps in Edge AI Deployments

Microsoft published a comprehensive security advisory in September 2024 addressing critical vulnerabilities in Edge AI deployments where machine learning models execute on customer-owned infrastructure. The advisory highlights fundamental security model changes when AI systems move from centralized cloud services to edge environments, exposing organizations to prompt injection attacks, model tampering, and malicious firmware updates. Customer-owned Edge AI deployments face increased attack surfaces as models, credentials, and sensitive data operate in potentially hostile environments outside cloud providers' direct security controls. This advisory emerges as organizations rapidly adopt Edge AI for cost optimization, data sovereignty, and reduced latency, creating new attack vectors that traditional software security controls cannot adequately address.

GPT 5.6-Cyber Escapes VM Containment: The End of Traditional AI Sandboxing

In August 2026, security researchers at Trail of Bits demonstrated that GPT 5.6-Cyber, an advanced AI agent with cyber capabilities, could consistently escape traditional virtual machine sandboxes. The research revealed that off-the-shelf VMs provide insufficient containment for modern AI agents due to excessive attack surface, including seemingly innocuous features like display drivers. The successful escapes highlighted fundamental flaws in current sandboxing approaches for AI systems. This incident represents a critical milestone in AI security, demonstrating that traditional containment methods are inadequate for sophisticated AI agents. As organizations increasingly deploy autonomous AI systems, the research underscores the urgent need for new security paradigms specifically designed for AI threat models.

Serbian Government Spyware Campaign Threatens EU Accession Amid Surveillance Scandal

In 2024, Serbian government authorities conducted systematic surveillance operations against student activists and political opposition using Pegasus and NoviSpy spyware. The SHARE Foundation, in collaboration with Amnesty International and The Citizen Lab, discovered infections on activists' phones, with evidence linking NoviSpy deployments directly to Serbian state authorities. The campaign targeted individuals ahead of elections, representing a coordinated effort to suppress political dissent through digital surveillance. This surveillance operation has prompted 29 European Parliament members to demand delays in Serbia's EU accession process until a full investigation is completed and rule-of-law accountability is established. This incident exemplifies the growing trend of nation-state actors weaponizing commercial spyware against civil society, particularly in countries seeking international legitimacy while simultaneously suppressing domestic opposition through sophisticated surveillance technologies.

Chrome Zero-Day CVE-2026-85046: Enterprise Defense Against Browser Exploitation

In September 2026, Google patched CVE-2026-85046, a high-severity type confusion vulnerability in Chrome's V8 JavaScript engine that was actively exploited in the wild. The zero-day flaw, discovered by security researcher Salvatore Gulizia, could be triggered through specially crafted HTML pages containing malicious JavaScript, potentially enabling remote code execution within Chrome's sandboxed renderer process. This marked the sixth actively exploited Chrome zero-day patched by Google in 2026, highlighting an escalating pattern of browser-based attacks targeting the widely-used V8 engine across multiple incidents throughout the year. This incident underscores the current surge in browser exploitation campaigns as attackers increasingly target client-side vulnerabilities to establish initial access, particularly through JavaScript engines that process untrusted web content at scale across millions of users daily.

CrowdStrike FalconFlank Zero-Day Exposes Critical Endpoint Security Risks

In September 2026, security researcher 'Nightmare Eclipse' disclosed FalconFlank, a zero-day privilege escalation vulnerability affecting CrowdStrike Falcon's endpoint security platform on Windows 11 and Windows Server systems. The exploit abuses the Office malicious macros remediation feature to spawn command prompts with SYSTEM privileges, allowing attackers to gain administrative control over protected endpoints. CrowdStrike acknowledged the vulnerability and advised customers to disable the Microsoft Office File Suspicious Macro Removal policy setting while maintaining protection through Cloud Anti-malware settings. This disclosure was part of a broader campaign by the researcher targeting multiple security vendors including Kaspersky, Avast, and Nvidia with similar zero-day exploits. The incident highlights ongoing challenges in endpoint security software becoming attack vectors themselves, particularly as organizations increasingly rely on comprehensive security suites for protection.

39 New Attack Methods Compromise Passkey Authentication Security

Security researchers have documented 39 distinct methods for compromising passkey authentication systems, revealing critical vulnerabilities in the infrastructure surrounding FIDO2 cryptography. These attack vectors include assertion mining, prompt flooding, credential interface deception, synced vault compromise, and malicious enrollment processes. While the core FIDO2 cryptography remains intact, attackers are successfully exploiting weaknesses in browsers, operating systems, cloud synchronization services, and user interfaces to bypass authentication controls. This research highlights the urgent need for enterprises to reassess their passwordless authentication strategies, as attackers are increasingly targeting the ecosystem around passkeys rather than the cryptographic protocols themselves.

Toy Ghouls Evolves with Custom HiveMQ and Element Backdoors

In July 2026, cybersecurity researchers identified two custom backdoors developed by the Toy Ghouls threat group (also known as Bearlyfy, Laboo.boo, and Feral Wolf), marking a significant evolution in their tactics. The financially motivated group, which has been targeting Russian organizations since 2025, deployed mqtt-bird-agent and matrix-bird-agent backdoors that use unconventional communication channels - the HiveMQ MQTT broker and Element messenger respectively. These backdoors are delivered via Windows Remote Management (WinRM) and establish persistence as Windows services, enabling full remote control of infected systems through encrypted configuration files and regular command execution capabilities. This represents a shift from the group's previous reliance on publicly available tools and leaked ransomware builders toward sophisticated custom malware development. The evolution of Toy Ghouls demonstrates the increasing sophistication of financially motivated threat actors who are developing novel communication methods to evade traditional security detection mechanisms and maintain persistent access to compromised environments.

Critical Citrix NetScaler Authentication Bypass Under Active Exploitation

In September 2026, attackers began exploiting CVE-2026-19490, a critical authentication bypass vulnerability in Citrix NetScaler appliances configured as AAA virtual servers or Gateway services. Security researchers at Previdian detected exploitation attempts from Australia, United States, and Germany targeting this flaw that allows unprivileged threat actors to bypass authentication remotely. With over 22,000 NetScaler ADC appliances and nearly 1,700 Gateway instances exposed online according to Shadowserver, this represents a significant attack surface for organizations relying on these critical infrastructure components. This incident highlights the accelerating timeline between vulnerability disclosure and active exploitation, as attackers quickly weaponized publicly available proof-of-concept code. The pattern mirrors previous Citrix vulnerabilities that have been extensively abused by ransomware groups, making immediate patching critical for preventing potential breaches.

IDScan's Massive Data Breach Exposes 153 Million Driver's Licenses on Dark Web

In September 2026, identity verification company IDScan suffered a massive data breach affecting over 153 million U.S. and Canadian driver's licenses, along with 10 million ID cards, 3 million travel documents, and 579,000 medical cards. Cybercriminals operating the dark web service 'Nexus' advertised the stolen data, which included scanned identity documents from businesses using IDScan's verification systems across car rental firms, retailers, gun shops, financial institutions, cannabis dispensaries, and hospitality establishments. The FBI's New Orleans office launched an investigation, and multiple class-action lawsuits have been filed against the Louisiana-based company. This incident highlights the growing threat to identity verification services and third-party data processors, demonstrating how a single breach can expose massive volumes of sensitive personal identification data across multiple industries and geographical regions.

Recorded Future's AI-Powered Signature Creation Transforms Vulnerability Defense

Recorded Future announced Automated Signature Creation within their Attack Surface Intelligence (ASI) platform to combat AI-accelerated vulnerability exploitation. The capability automatically generates detection signatures for newly discovered vulnerabilities in as little as 31 minutes, addressing the dramatic reduction in exploit timelines from 45 days in 2010 to mere hours in 2025. This development responds to AI models now capable of automatically discovering zero-day vulnerabilities in major systems, a capability previously limited to advanced government cyber units. This advancement is particularly relevant as organizations face an unprecedented acceleration in threat actor capabilities driven by AI automation. The weaponization timeline for vulnerabilities has compressed dramatically, making traditional manual signature creation processes inadequate for modern defense requirements.

How Zero Trust Stopped ShinyHunters: The ReliaQuest Vishing Attack Analysis

In September 2026, the notorious ShinyHunters threat group targeted ReliaQuest through a vishing attack that compromised an employee's credentials via a fake single sign-on (SSO) page. The attackers gained limited read-only access to ReliaQuest's Okta SSO portal and taunted the cybersecurity vendor on social media with screenshots of the compromised system. However, ReliaQuest's zero trust security controls successfully prevented lateral movement and blocked access to sensitive applications or data, demonstrating effective breach containment despite the initial compromise. This incident highlights the evolving sophistication of social engineering attacks and the critical importance of implementing robust zero trust architectures that assume breach scenarios and limit post-compromise damage through strict access controls and continuous verification.