audio-separator versions before 0.47.2 have an insecure deserialization vulnerability in the handling of Demucs .th model files. Attackers able to supply malicious model files loaded via torch.load with weights_only=False can execute arbitrary code. This affects local users who can write to the shared /tmp/audio-separator-models/ directory or attackers controlling unverified model download sources. The vulnerability is fixed in version 0.47.2.
Red Hat has issued a security update for Red Hat Hardened Images RPMs addressing a critical vulnerability identified as CVE-2026-114938 in the java-21-openjdk-portable package. The update includes multiple RPMs for various architectures such as aarch64, s390x, and x86_64. The advisory does not provide detailed technical information about the vulnerability but classifies it as critical and associates it with CWE-787 (Out-of-bounds Write).
CVE-2026-133937 is a vulnerability in highlight.js up to version 11.12.0 affecting the hljs.highlight function in the livescript language grammar handler. It involves inefficient regular expression complexity triggered by crafted input, which can be exploited remotely. An exploit has been published, but a fix is pending acceptance of a pull request.
Unit 42 has analyzed a series of vishing attacks and extortion campaigns by Com-affiliated threat actors targeting data-rich sectors such as healthcare, finance, and legal services primarily in the U.S., UK, and Australia. These attacks involve social engineering to gain initial access, use of adversary-in-the-middle proxies to capture credentials and MFA approvals, automated data exfiltration from Microsoft 365 environments, and extortion via internal messaging directing victims to Tor-hosted negotiation sites. A malicious browser extension masquerading as Google Chrome was also identified, capable of keylogging, stealing authentication cookies, capturing webcam feeds, and remote control via Chrome DevTools Protocol. Multiple named threat clusters and actors are involved, with some newly identified in late 2026. The activity is ongoing with detailed telemetry and incident response engagements informing the analysis.
The signed-manifest-updater package on PyPI (version 0.1.5) is part of a malicious campaign that delivers remote access trojans (RATs). It overrides the installation process to download and execute a remote script, which is malicious in intent. Although the analyzed code was benign at the time, the package is linked to a campaign known for infostealing malware.
OceanWP WordPress Theme versions before 4.2.9 contain a stored cross-site scripting (XSS) vulnerability in the oceanwp_edd_terms_list() function. Authenticated users with Easy Digital Downloads (EDD) download category creation permissions can inject malicious JavaScript by crafting a specially formed category name. This payload executes in the browsers of visitors who interact with the affected category link on the public EDD archive page, potentially compromising administrator sessions if they view the page while authenticated. The vulnerability is fixed in version 4.2.9.
A vulnerability in continuedev continue up to version 2.1.0 affects the DataLogger.logDevData function in the Dev Data Logging component, causing information disclosure. The flaw can be exploited remotely and requires user interaction. The vendor has not responded to the disclosure. The CVSS score is 4.3, indicating medium severity.
A server-side request forgery (SSRF) vulnerability exists in continuedev continue versions 2.0.0 and 2.1.0 within the DocsCrawler component. The vulnerability arises from manipulation of the startUrl or useLocalCrawling arguments in the DocsCrawler.ts file. Exploitation requires a high level of complexity and user interaction, and is considered difficult. No vendor response or patch is currently available.
The WPC Smart Messages for WooCommerce plugin versions before 4.3.6 have a missing authorization vulnerability. Authenticated users with Contributor or Author roles can exploit this flaw by using specially crafted shortcodes to access arbitrary post meta and private product data without proper capability checks. This allows exposure of sensitive customer and business information such as billing emails, phone numbers, addresses, private product descriptions, and purchase notes to unauthenticated visitors.
CVE-2026-133697 is a low-severity information disclosure vulnerability in the continuedev continue software up to version 2.1.0. It affects the LocalPlatformClient.resolveFQSNs function in the core/config/yaml/LocalPlatformClient.ts file within the Config Resolution component. Exploitation requires local access and manipulation of the Name argument. The vendor has not responded to disclosure attempts, and no patch or fix information is currently available.
A security vulnerability (CVE-2026-133694) exists in continuedev continue versions 2.0.0 and 2.1.0. The flaw is in the getSidebarContent function within the Mermaid Diagram Renderer component, specifically in the file extensions/vscode/src/ContinueGUIWebviewViewProvider.ts. This vulnerability allows remote attackers to perform OS command injection. The vendor was contacted but has not responded or provided a fix. The CVSS score is 6.3, indicating a medium severity risk.
CVE-2026-19777 is a denial-of-service vulnerability in the G DATA Total Security AVKProxy component. It allows a local attacker with low-privileged code execution to create a symbolic link that abuses the G DATA ANTIVIRUS Proxy Service to cause a persistent denial-of-service condition.
A heap-based buffer overflow vulnerability in Open Robotics RViz allows network-adjacent attackers to execute arbitrary code without authentication. The flaw is in the handling of mesh resources in PointCloud and Marker displays, where user-supplied data length is not properly validated before copying to a heap buffer.
A denial-of-service vulnerability exists in Open Robotics RViz's RGB8PCTransformer component due to an out-of-bounds read in PointField handling. This flaw allows network-adjacent attackers to cause a crash without authentication by exploiting improper validation of PointCloud2 message data.
A denial-of-service vulnerability exists in Open Robotics RViz due to improper validation of OccupancyGridUpdate messages, allowing out-of-bounds reads. Network-adjacent attackers can exploit this without authentication to cause a denial-of-service condition.
CVE-2026-19779 is a local privilege escalation vulnerability in the Backup Service component of G DATA Total Security. It arises from an uncontrolled search path element where the service restores archives from an unsecured location. Exploitation requires an attacker to already have the ability to execute low-privileged code on the system. Successful exploitation allows arbitrary code execution with SYSTEM privileges, leading to full system compromise.
A local privilege escalation vulnerability exists in the TSNxGService component of G DATA Total Security. An attacker with the ability to execute low-privileged code can exploit this flaw by creating a symbolic link that causes the service to delete an arbitrary file. Successful exploitation allows escalation to SYSTEM privileges and arbitrary code execution.
A denial-of-service vulnerability exists in Open Robotics RViz due to an out-of-bounds read in the XYZPCTransformer component when processing PointCloud2 messages. This flaw allows network-adjacent attackers to cause a crash without authentication. The vulnerability arises from improper validation of user-supplied data, leading to a read past the allocated memory boundary.
A denial-of-service vulnerability exists in Open Robotics RViz's AxisColorPCTransformer component due to an out-of-bounds read in handling PointCloud2 messages. This flaw allows network-adjacent attackers to cause a denial-of-service without requiring authentication.