All news with #privilege escalation tag

⚠️ XM Cyber disclosed privilege-escalation flaws in Google’s Vertex AI that let low‑privileged users manipulate Google-managed Service Agents to gain elevated project-wide permissions. Google told XM Cyber this behavior is "working as intended." Security experts warn that managed service identities and insecure defaults create invisible, structural risks. CISOs are urged to audit service identities, reduce authentication scope, and monitor agent activity like privileged users.

🔓 Modular DS versions 2.5.1 and earlier contain a critical privilege-escalation bug (CVE-2026-23550) that lets unauthenticated attackers gain full WordPress admin access by calling unprotected API routes under /api/modular-connector/. Patchstack reported active exploitation and the vendor released Modular DS 2.5.2 on January 14, 2026. Administrators should update immediately, check for rogue admin accounts, enable two-factor authentication, apply IP restrictions, and consider Patchstack’s mitigation rules if immediate patching isn’t possible.

⚠️ Patchstack reports a maximum-severity vulnerability (CVE-2026-23550, CVSS 10.0) in the Modular DS WordPress plugin affecting all versions up to and including 2.5.1. The flaw permits unauthenticated privilege escalation via routes under /api/modular-connector/ when the "direct request" mode with an "origin=mo" parameter is used, bypassing authentication. Exploitation was observed beginning Jan 13, 2026, and the issue is patched in 2.5.2; administrators should update immediately.

⚠️ AVEVA has released patches for multiple vulnerabilities in Process Optimization that could allow remote code execution, SQL injection, privilege escalation, and disclosure of sensitive data. The most severe, CVE-2025-61937, permits unauthenticated remote code execution at OS System privileges (CVSS 10.0). AVEVA's remediation requires updating to Process Optimization v2025; CISA and the vendor also recommend firewall restrictions, ACLs, and ensuring encrypted channels.

🔒 Organizational AI agents are increasingly embedded in critical workflows and often run under shared service identities with broad, long-lived permissions. Because actions execute under the agent identity, users can indirectly obtain access they don’t have, and audit logs typically attribute activity to the agent rather than the initiating user. This creates invisible privilege-escalation paths and complicates least-privilege enforcement. Wing is cited for continuously discovering agents, mapping their access to critical assets, and restoring visibility and accountability.

⚠ Siemens disclosed a local privilege escalation vulnerability (CVE-2025-40942) in TeleControl Server Basic affecting product versions earlier than V3.1.2.4. The flaw could allow an attacker with local access to execute arbitrary code with elevated privileges and is rated High under CVSS 3.1 (8.8). Siemens released V3.1.2.4 to remediate the issue. Administrators should apply the update promptly and follow network-segmentation and access-control best practices to reduce exposure.

🔒 Microsoft released the KB5073724 Extended Security Update for Windows 10, available to Windows 10 Enterprise LTSC and systems enrolled in the ESU program. Install via Settings → Windows Update by performing a manual “Check for Updates”; installs update and raises builds to 19045.6809 (Windows 10) and 19044.6809 (Enterprise LTSC 2021). The update contains only security and bug fixes — including patches for three zero-days, an actively exploited elevation-of-privilege fix in Agere modem drivers, an updated WinSqlite3.dll, and targeted handling for expiring Secure Boot certificates.

🛡️ Trend Micro has released a security update for Apex Central after vulnerability management vendor Tenable identified multiple serious flaws affecting all on-premises builds earlier than 7190. The most severe is a 9.8-rated LoadLibraryEX issue that can allow an unauthenticated attacker to force the server to load and execute an attacker-controlled DLL as SYSTEM. Two additional high-severity, unauthenticated flaws can cause denial-of-service. Trend Micro urges customers to apply build 7190 and review remote access controls immediately.

🔒 Researchers disclosed 11 critical vulnerabilities in Coolify, an open-source self-hosting platform, including multiple authenticated command injections, remote code execution, container escape and an information disclosure of the root SSH private key. Several issues carry CVSS scores of 9.4–10.0 and allow attackers with low or moderate privileges to execute arbitrary commands as root or obtain persistent access. Operators should upgrade to patched releases or apply vendor mitigations immediately.

🔒 Veeam has released a patch addressing four vulnerabilities in Backup & Replication v13 that let users with Backup Admin, Backup Operator, or Tape Operator roles exceed intended privileges. The most severe, CVE-2025-59470 (CVSS 9.0), can enable remote code execution as the Postgres user; others permit file writes as root or RCE via malicious configuration files. Veeam recommends immediate installation of version 13.0.1.1071; the vendor says core backup data remains immutable and intact.

🔒 Kaspersky observed Mustang Panda leveraging a signed, previously undocumented kernel‑mode rootkit driver to deliver a new TONESHELL backdoor in mid‑2025 against targets in Asia. The driver, tracked as ProjectConfiguration.sys, uses an old certificate issued to Guangzhou Kingteller Technology Co., Ltd., likely leaked or stolen, and registers as a high‑altitude minifilter to intercept I/O. It spawns an injected svchost.exe and loads a memory‑only TONESHELL implant that communicates with C2 servers and resists disk‑based detection.

🐀 Attackers are hosting counterfeit proof-of-concept exploit repositories on GitHub to deliver the Webrat backdoor to unsuspecting users. Kaspersky analysts observed polished, likely machine-generated README files that mask a password-protected ZIP; the archive password is hidden in filenames and often missed. Inside are decoy DLLs, batch loaders and executables (e.g., rasmanesc.exe) that disable Windows Defender, escalate privileges, and fetch the real payload from hardcoded C2 servers. The campaign, active since at least September 2025, appears tuned to catch novice researchers and students who analyze PoCs outside isolated environments.

🛡️ Kaspersky researchers found WebRAT backdoor being distributed through GitHub repositories that posed as proof‑of‑concept exploits for recently disclosed vulnerabilities. The malicious packages were delivered as password‑protected ZIPs containing a corrupted decoy DLL, a batch script, and a main dropper named rasmanesc.exe that elevates privileges, disables Defender, and downloads WebRAT. All identified repositories have been removed, but developers are urged to verify PoC sources and test untrusted code in isolated environments.

⚠️ Certain motherboard models from vendors including ASRock, ASUS, GIGABYTE, and MSI are affected by a firmware flaw that reports DMA protection as active but fails to initialize the IOMMU during early boot. That discrepancy allows a physically present attacker with a DMA-capable PCIe device to read or modify system memory and potentially enable pre-boot code injection before OS protections load. CERT/CC warned the gap undermines boot integrity and access to sensitive memory. Affected vendors have released firmware updates to correct the IOMMU initialization sequence; users and administrators should apply patches promptly.

⚠️ Inductive Automation Ignition contains a Python scripting vulnerability (CVE-2025-13911) that can allow direct SYSTEM-level code execution on Windows hosts running the Ignition Gateway. The issue stems from insufficient controls on which Python libraries and scripts can be imported and executed, and the Ignition service account running with excessive SYSTEM privileges. A malicious project uploaded by an authenticated administrator can execute bind shells or similar payloads with Gateway process privileges. Inductive Automation identifies affected releases as 8.1.x and 8.3.x and provides mitigations on its Trust Portal; CISA rates the flaw CVSS 3.1 6.4 and recommends network segmentation and reduced exposure.

⚠ SonicWall released patches addressing CVE-2025-40602 (CVSS 6.6), a local privilege escalation in the Secure Mobile Access (SMA) 100 Appliance Management Console caused by insufficient authorization. Affected firmware builds prior to 12.4.3-03245 and 12.5.0-02283 have updates available to remediate the issue. SonicWall said the flaw has been actively exploited and has been observed chained with CVE-2025-23006 to achieve unauthenticated remote code execution with root privileges; users should apply fixes immediately.

⚠️ SonicWall warns of a medium-severity local privilege escalation in the SonicWall SMA1000 Appliance Management Console (CVE-2025-40602), reported by Google Threat Intelligence researchers Clément Lecigne and Zander Work. The vendor says this LPE was chained in active zero-day attacks with a critical pre-auth deserialization bug (CVE-2025-23006) to execute OS commands and escalate to root. Administrators should apply the vendor hotfix and firmware updates immediately.

🔓 A critical arbitrary file upload vulnerability in the Motors WordPress theme could let low-privileged, logged-in users install and activate plugins, enabling remote code execution and full site takeover. The flaw, tracked as CVE-2025-64374, affects versions 5.6.81 and earlier and was discovered by Denver Jackson of the Patchstack Alliance community. The issue stems from an AJAX handler that relies on a nonce for validation but lacks a proper permission check, allowing Subscriber-level users to supply arbitrary plugin URLs. The vendor released a fix in version 5.6.82 on 3 November; site owners should update immediately to mitigate the risk.

⚠️ The JumpCloud Remote Assist for Windows agent contains a critical local privilege escalation vulnerability (CVE-2025-34352) that can be exploited during uninstall or update flows. The uninstaller runs with NT AUTHORITY\SYSTEM and performs file operations in a user-writable %TEMP% subdirectory without validating or securing the path. Attackers with a local foothold can abuse link-following techniques (mount points and symlinks) to overwrite or delete protected files, yielding full system compromise or denial-of-service. Systems running Remote Assist before version 0.317.0 should be updated immediately.

⚠️ Security researchers have identified a critical vulnerability (CVE-2025-34352) in the JumpCloud Remote Assist Windows agent that allows low-privileged local users to escalate to NT AUTHORITY\SYSTEM or trigger denial-of-service during uninstallation. The root cause is unsafe file operations in user-writable directories (notably %TEMP%), enabling link-following attacks that redirect privileged actions. XM Cyber reported the issue and JumpCloud has released version 0.317.0 to address it — administrators should update affected endpoints immediately.