Security researchers at Oasis Security have uncovered a significant vulnerability in Microsoft’s OneDrive File Picker, revealing that websites using the tool can gain access to a user’s entire OneDrive storage not just the specific files intended for upload. This flaw, which affects hundreds of apps including ChatGPT, Slack, Trello, and ClickUp, could impact millions of … Continued
GreyNoise has discovered a sophisticated campaign exploiting ASUS routers exposed to the internet,using CVE-2023-39780 (command injection vulnerability). This operation is not just a one-off attack, but part of a broader effort to construct a distributed network of compromised devices, likely intended for future botnet operations. The attackers behind this campaign are not your average opportunists. … Continued
Sophos recently uncovered a targeted cyberattack on a Managed Service Provider (MSP) after detecting suspicious activity involving the MSP’s remote monitoring and management (RMM) tool SimpleHelp. The investigation revealed that a threat actor had compromised the RMM platform and used it to deploy DragonForce ransomware across multiple customer environments, exfiltrate sensitive data, and launch a … Continued
In a statement published on its official website, Adidas has confirmed a data breach involving customer information accessed by an “unauthorised external party.” The sportswear giant clarified that the incident occurred not within its own systems, but through a third-party customer service provider. According to Adidas, the exposed information may include: Full names Email addresses … Continued
Microsoft has released an out-of-band (OOB) emergency update to address a critical issue causing some Hyper-V virtual machines (VMs) running on Windows Server 2022 to freeze or restart unexpectedly. The problem primarily affects Azure Confidential VMs, a specialized class of virtual machines designed to secure data during processing, transmission, and storage. The issue stems from … Continued
Invariant has discovered a critical vulnerability in GitHub’s Model-Centric Programming (MCP) integration—one that allows attackers to hijack AI agents using a malicious GitHub issue and leak sensitive data from private repositories. This vulnerability demonstrates a growing concern in the intersection of AI agents and developer workflows: indirect prompt injection, or what invariant calls them toxic … Continued
Socket’s Threat Research Team has identified a coordinated and ongoing malicious campaign targeting the npm ecosystem, involving at least 60 packages designed to quietly exfiltrate sensitive system information to a threat actor-controlled Discord webhook. What Does it Do? The malicious packages published under three separate npm accounts deploy post-install scripts that activate during npm install … Continued
A bipartisan group of lawmakers is making a renewed effort to bolster cybersecurity requirements for federal government contractors. Senators Mark Warner (D-Va.) and James Lankford (R-Okla.) have reintroduced the Federal Contractor Cybersecurity Vulnerability Reduction Act, which would require contractors to adhere to guidelines set by the National Institute of Standards and Technology (NIST) for vulnerability … Continued
Just last week, the fictional Ethan Hunt outwitted a rogue AI to save world in Mission: Impossible. But in real life, a new set of experiments from Palisade Research is raising serious concerns about how some AI systems are beginning to act in ways that directly defy human instructions. Palisade Research recently made waves with … Continued
A newly disclosed vulnerability in ModSecurity2 firewall, tracked as CVE-2025-47947, has raised concerns over potential Denial of Service (DoS) attacks under specific, rare conditions. The issue was officially published on May 21, 2025, and is rated 7.5 (High) on the CVSS scale. The vulnerability was initially reported privately by a customer in March 2025. After … Continued
