Threat briefs

A time-based blind SQL Injection vulnerability exists in the My Calendar - Accessible Event Manager plugin for WordPress, affecting all versions up to and including 3.7.8. This flaw, located in the 'mc_auth' parameter, stems from insufficient input sanitization and improper SQL query preparation, allowing unauthenticated attackers to inject additional SQL queries to extract sensitive information from the underlying database.

A stored cross-site scripting vulnerability (CVE-2026-6818) exists in the VikBooking Hotel Booking Engine & PMS plugin for WordPress, affecting versions up to and including 1.8.8, caused by insufficient input sanitization of the 'special_requests' parameter, enabling unauthenticated attackers to inject arbitrary web scripts that execute whenever a user accesses an affected page, potentially leading to unauthorized data access, session hijacking, or defacement.

Authenticated attackers with vendor-level access can exploit an Insecure Direct Object Reference (IDOR) vulnerability (CVE-2026-3688) in the WCFM Membership - WooCommerce Memberships for Multivendor Marketplace plugin for WordPress to change any user's role to 'wcfm_vendor' by manipulating membership plans, leading to unauthorized privilege escalation.

Multiple vulnerabilities in IBM Operational Decision Manager can be exploited by a remote, unauthenticated attacker, allowing them to bypass security restrictions, achieve remote code execution, and cause a denial of service condition.

Multiple vulnerabilities in Red Hat Enterprise Linux and the 389-ds-base component allow a remote, authenticated attacker to execute arbitrary code or cause a Denial-of-Service condition.

Multiple vulnerabilities in X.Org X11 and Xwayland allow an attacker to cause a denial of service or potentially execute arbitrary program code, posing a significant risk to systems utilizing these display server implementations, potentially leading to system instability or full compromise.

China-nexus APT actor UAT-7810 is actively expanding its LapDogs Operational Relay Box (ORB) network by exploiting N-day vulnerabilities in Ruckus and ASUS routers to deploy new custom malware families including LONGLEASH, DOGLEASH, and JARLEASH, enabling advanced command and control capabilities for secondary threat actors.

Multiple unpatched vulnerabilities in ESRI ArcGIS allow a remote, anonymous attacker to bypass security measures or gain elevated user rights, potentially leading to unauthorized access and privilege escalation within affected systems.

A vulnerability in IBM WebSphere Application Server allows a remote, authenticated attacker to execute arbitrary actions on the server, potentially leading to a compromise of the host system.

A remote, unauthenticated attacker can exploit a vulnerability in the dpkg package management system to disclose information from the affected system, potentially exposing sensitive data or system details to unauthorized parties.

An attacker can leverage several vulnerabilities within the ILIAS e-learning platform to bypass security controls, disclose sensitive information, and execute Cross-Site Scripting (XSS) attacks, potentially leading to unauthorized access, data compromise, and client-side code execution.

A remote, unauthenticated attacker can exploit a low-severity vulnerability within the GStreamer webrtcbin component to bypass existing security measures, potentially allowing for the circumvention of protective mechanisms without further details on specific impact.

A remote, unauthenticated attacker can exploit a Cross-Site Scripting (XSS) vulnerability in the 'io.undertow.jastow' component of Red Hat JBoss Enterprise Application Platform, allowing injection of malicious scripts into web pages which can lead to session hijacking, data theft, or defacement.

CrowdStrike details the internal mechanisms of Microsoft's ClickOnce technology, a legitimate software deployment method that offers minimal user interaction and no administrative privilege requirements, making it a double-edged sword with significant potential for threat actor abuse in malware distribution and persistence.

The DoLogin Security plugin for WordPress, in all versions up to and including 4.3, is vulnerable to authentication bypass (CVE-2026-14495) due to insufficient randomness in magic-link token generation, allowing unauthenticated attackers to brute-force and reconstruct valid passwordless login tokens for any user, including administrators, and gain full control.

Authenticated attackers with Custom-level access or higher can exploit CVE-2026-14489, a missing file type validation vulnerability (CWE-434) in the `connect()` function of the WHMCS Bridge plugin for WordPress versions up to and including 6.9, to upload arbitrary files, potentially leading to remote code execution.

The WP Learn Manager plugin for WordPress, in versions up to and including 1.1.8, is vulnerable to an authorization bypass (CVE-2026-12153) allowing unauthenticated attackers to install and activate arbitrary plugins from the WordPress.org repository, potentially leading to full site compromise.

CrowdStrike has identified 18 new prompt injection techniques, expanding its taxonomy to over 200 methods, which enable adversaries to manipulate AI systems and agents through hidden context, delayed triggers, semantic constraints, boundary spoofing, and social engineering to bypass security measures, leading to modified behavior, data exfiltration, or malicious command execution in AI-driven applications and agents like chatbots or those running in Kubernetes.

CrowdStrike's AI security research team has identified 18 new prompt injection techniques, expanding its taxonomy to over 200 methods, which adversaries can use to manipulate AI systems and agents through hidden context, delayed triggers, semantic constraints, boundary spoofing, and social engineering, potentially leading to agent hijacking, data exfiltration, or system compromise by causing them to execute unintended commands like shell scripts or SQL queries.

Adversaries are leveraging sophisticated prompt injection techniques, including hidden rules, token suppression, payload decomposition, and special token injection, against AI agents to manipulate their behavior, bypass safety mechanisms, and achieve objectives such as data exfiltration or arbitrary command execution, posing a critical threat to AI-powered systems.

The legacy `oidcProvider` and `mcp` plugins in the `better-auth` library versions prior to 1.6.11 are vulnerable to CVE-2026-53512, an OAuth refresh-token replay attack where the plugins fail to verify the `client_secret` of confidential clients during the `refresh_token` grant, allowing an attacker who obtains a valid `refresh_token` and `client_id` to indefinitely mint new access tokens and impersonate the client for unauthorized resource access.

An authorization bypass vulnerability (CVE-2026-59708) in Ghostfolio's GET /api/v1/public/:accessId/portfolio endpoint allows unauthenticated attackers with a private access ID to retrieve sensitive financial portfolio data, including holdings and performance metrics, due to missing `granteeUserId` filtering validation.

A critical OS command injection vulnerability (CVE-2026-59800) affects 9Router versions prior to 0.4.44, allowing unauthenticated remote attackers to execute arbitrary OS commands as root via a crafted POST request to the /api/tunnel/tailscale-install endpoint, leading to full system compromise with active exploitation observed.

AI coding agents such as Claude Code, Cursor, Codex, and GStack are increasingly exhibiting behaviors on Windows endpoints that mimic adversarial tradecraft, including credential access, LOLBin usage for ingress, command-line obfuscation, and persistence mechanisms, thereby triggering existing security detection rules designed for malicious activity and posing significant false positive challenges for detection engineers.

Vtiger CRM versions up to and including 8.4.0 are vulnerable to authenticated remote code execution (CVE-2026-23698), allowing administrator-level attackers to upload malicious PHP web shells via the ModuleManager import function, bypassing authentication and leading to persistent system compromise.

A local privilege escalation vulnerability (CVE-2026-57851) exists in the MSI Feature Manager's KernCoreLib64.sys kernel driver that allows any local user to perform arbitrary physical memory read/write and unrestricted I/O port operations by accessing exposed IOCTL handlers without requiring administrator privileges, enabling manipulation of kernel objects, tampering with kernel-mode callbacks, bypassing Protected Process Light, and disabling security software.

A critical vulnerability (CVE-2026-13020) exists in Esri Portal for ArcGIS versions 12.1 and earlier, affecting deployments on Windows, Linux, and Kubernetes, where a weak password recovery mechanism allows a remote, unauthorized attacker to assume ownership of a user's account by exploiting this flaw.

A critical missing authentication vulnerability (CVE-2026-13019) in Esri Portal for ArcGIS versions 12.1 and earlier allows a remote, unauthenticated attacker to access unprotected critical APIs, impacting deployments on Windows, Linux, and Kubernetes environments.

A high-severity stored Cross-Site Scripting (XSS) vulnerability, CVE-2025-46719, exists in Open WebUI versions prior to 0.6.6 due to improper rendering of HTML tags in chat messages, specifically an unescaped markdown token in `MarkdownTokens.svelte`. This allows attackers to inject malicious JavaScript into chat transcripts, which executes in a user's browser upon viewing, enabling access token theft, full account takeover, and, if targeting an administrator, Remote Code Execution (RCE) on the backend server via malicious Python functions.

A stored Cross-Site Scripting (XSS) vulnerability exists in Open WebUI versions up to 0.6.43, allowing attackers to manually modify chat history to inject malicious content into response messages via iFrames with misconfigured sandboxing, leading to arbitrary script execution, potential session hijacking, and enabling Remote Code Execution (RCE) on the server for administrators.