All news in category “Security Advisory and Patch Watch”

⚠ Security researchers at Cato Networks disclosed CVE-2025-64496, a vulnerability in Open WebUI that lets external model servers inject JavaScript via Server-Sent Events (SSE) when the Direct Connections feature is enabled. An attacker controlling a malicious model endpoint can exfiltrate JSON Web Tokens (JWTs) from the browser, enabling account takeover and access to documents, chats, and embedded API keys. If the compromised account has Workspace Tools privileges, the session token can be used to execute authenticated Python code on the backend, leading to remote code execution. The flaw affects versions up to 0.6.34 and is fixed in 0.6.35; organizations are urged to update and implement HttpOnly cookies, strict CSPs, and ban dynamic code evaluation.

⚠️ A maximum-severity flaw, CVE-2026-21858 (Ni8mare), in n8n allows unauthenticated remote attackers to read local files, forge administrator sessions, and achieve remote code execution by exploiting a Content-Type parsing confusion that can override req.body.files. The bug affects releases up to and including 1.65.0 and was fixed in 1.121.0 (released November 18, 2025). Operators should upgrade immediately, avoid exposing n8n publicly, and restrict or disable public webhooks and form endpoints until patched.

⚠️ Veeam released security updates for Backup & Replication to fix multiple vulnerabilities, including a remote code execution bug tracked as CVE-2025-59470. The flaw affects version 13.0.1.180 and earlier 13 builds and can allow users with Backup or Tape Operator roles to execute code as the postgres user. On January 6 Veeam published 13.0.1.1071 to patch CVE-2025-59470 plus a high (CVE-2025-55125) and a medium (CVE-2025-59468) issue. Administrators are advised to apply updates and follow Veeam's security guidelines to limit privileged-role exposure.

🔔 CISA has added two vulnerabilities to its Known Exploited Vulnerabilities (KEV) Catalog: CVE-2009-0556 (Microsoft Office PowerPoint code injection) and CVE-2025-37164 (HPE OneView code injection). CISA notes evidence of active exploitation and highlights that these vulnerability types are frequent attack vectors posing significant risk to the federal enterprise. Under BOD 22-01, Federal Civilian Executive Branch agencies are required to remediate KEV entries by the specified due date. CISA strongly urges all organizations to prioritize timely remediation as part of sound vulnerability management.

🔒 n8n has warned of a maximum-severity remote code execution flaw, CVE-2026-21877, rated 10.0 under CVSS. Under certain conditions an authenticated user may cause untrusted code to be executed by the service, potentially allowing full compromise of affected instances. Both self-hosted and n8n Cloud deployments running versions >= 0.123.0 and < 1.121.3 are impacted; the issue is fixed in 1.121.3 (released November 2025). Administrators should upgrade immediately or, if that is not possible, disable the Git node and restrict access for untrusted users.

🔒 Veeam has released security updates for Veeam Backup & Replication to address a critical remote code execution flaw tracked as CVE-2025-59470 (CVSS 9.0) that could allow a Backup or Tape Operator to run code as the postgres user via a crafted interval or order parameter. The vendor also fixed three additional vulnerabilities that permit escalation to root or file writes by privileged backup roles. All 13.x builds up to 13.0.1.180 are affected and the fixes are included in 13.0.1.1071; customers are advised to apply updates and follow role-hardening guidance promptly.

⚠️A critical remote code execution flaw, CVE-2026-0625, is being actively exploited in legacy D-Link DSL gateway routers via a command-injection weakness in the dnscfg.cgi endpoint. Improper sanitization of DNS configuration parameters allows unauthenticated attackers to execute arbitrary shell commands and modify DNS settings. D-Link says it is investigating affected firmware variants and will publish an updated model list after a firmware-level review. Owners of end-of-life devices should retire or replace impacted hardware immediately.

⚠An unauthenticated command injection (CVE-2026-0625) in dnscfg.cgi allows remote shell execution on multiple legacy D-Link DSL gateway routers. VulnCheck reported the issue to D-Link after The Shadowserver Foundation observed an exploitation attempt on a honeypot on December 15. Confirmed affected models (DSL-526B, DSL-2640B, DSL-2740R, DSL-2780B) are End-of-Life and will not receive patches. D-Link advises retiring affected devices or isolating them in segmented non-critical networks and applying restrictive security settings.

📧 Phishing actors are exploiting complex mail routing and misconfigured spoof protections to send messages that appear to originate internally, frequently using PhaaS platforms such as Tycoon2FA. Microsoft observed increased use of this vector since May 2025, including nested redirect chains and AiTM techniques to harvest credentials. Tenants with MX records pointed to Office 365 benefit from built-in protections; others must enforce strict SPF hard-fail, DKIM signing, and DMARC reject policies and correctly configure connectors to prevent these spoofing campaigns.

⚠ An unpatched firmware error in the TOTOLINK EX200 wireless range extender can cause the device to start an unauthenticated root-level telnet service when specific malformed firmware files are processed. CERT/CC (CVE-2025-65606) says exploitation requires an attacker to be authenticated to the web management interface to reach the firmware-upload handler, which can then enter an abnormal error state. The vendor has not issued a patch and the product is no longer actively maintained; users are advised to restrict administrative access and consider upgrading to a supported model.

⚠️ A high-severity vulnerability (CVE-2025-64496) affecting Open WebUI versions 0.6.34 and earlier can enable account takeover when the Direct Connections feature is enabled. A malicious OpenAI-compatible model server can send a crafted server-sent events message that executes JavaScript in a connected user's browser and steals authentication tokens from localStorage. Open WebUI 0.6.35 and later block the malicious execute events; administrators should upgrade immediately, restrict Direct Connections to trusted endpoints, and strengthen authentication and sandboxing.

⚠️ Columbia Weather Systems’ MicroServer firmware contains multiple vulnerabilities that could let an attacker redirect SSH connections, expose vendor and user secrets stored on an unencrypted SD card, and obtain a limited interactive shell with elevated file privileges. Affected devices run firmware versions prior to MS_4.1_14142. Columbia Weather Systems recommends updating to MS_4.1_14142 or later and contacting support for assistance; CISA advises minimizing network exposure, isolating control networks, and using secure remote access such as up-to-date VPNs. No known targeted public exploitation has been reported; UsrPacific reported these issues to CISA.

⚠️Security researchers disclosed a high-severity vulnerability in Open WebUI (CVE-2025-64496) that allows external model servers connected via the Direct Connections feature to stream server-sent events that execute JavaScript in the browser. Malicious code can read long-lived JSON Web Tokens stored in localStorage to take over accounts and access workspaces, documents, chats, and embedded API keys. With elevated workspace.tools permissions, attackers can escalate to remote code execution on backend servers. Organizations should patch to v0.6.35 immediately.

⚠️ A critical sandbox bypass, CVE-2025-68668 (CVSS 9.9), has been disclosed in n8n, allowing an authenticated user with workflow create/modify permissions to execute arbitrary OS commands on the host running n8n. The flaw resides in the Python Code Node that uses Pyodide and affects n8n versions 1.0.0 up to, but not including, 2.0.0. The issue is resolved in n8n 2.0.0, which makes the task-runner native Python implementation the default. Short-term mitigations include disabling the Code Node, disabling Python in the Code Node, or enabling the task-runner Python sandbox via environment variables.

🚨 Maintainers of @adonisjs/bodyparser urge immediate updates after disclosure of CVE-2026-21440, a critical path traversal flaw that can enable attackers to write arbitrary files via unsanitized multipart filenames. The vulnerability stems from MultipartFile.move(location, options) defaulting to client-supplied names when the options.name is omitted. Exploitation requires a reachable upload endpoint and can lead to file overwrite and possible RCE depending on deployment, filesystem permissions, and overwrite settings.

🔒 Bleeping Computer reports that attackers are actively exploiting an older FortiOS vulnerability, CVE-2020-12812, which can bypass two-factor authentication. Although Fortinet issued a patch in July 2020, researchers say at least 10,000 FortiGate firewalls remain unpatched. Administrators are urged to install the latest updates immediately to mitigate account access risks. Additional measures include restricting administrative access, rotating credentials, and monitoring logs for suspicious activity.

⚠ Administrators continue to find more than 10,000 internet-exposed Fortinet firewalls vulnerable to an active two-factor authentication bypass (CVE-2020-12812) that was patched in July 2020. The flaw in FortiOS SSL VPN permits login without a second factor when username case is altered; Fortinet advised disabling username case sensitivity as a mitigation. Shadowserver reports over 1,300 affected IPs in the U.S. — network owners should patch, apply mitigations, and audit LDAP-dependent management interfaces immediately.

🔒 IBM is urging customers to quickly apply interim fixes for a critical authentication-bypass vulnerability in IBM API Connect (CVE-2025-13915) that affects versions 10.0.8.0–10.0.8.5 and 10.0.11.0. The flaw can allow unauthorized access to exposed applications without user interaction and stems from a broken architectural assumption that traffic passing the gateway guarantees identity enforcement (CWE-305). IBM has published platform-specific interim fixes and advises disabling self-service sign-up on Developer Portals if patches cannot be applied; administrators must also remove image overrides when upgrading to avoid persistent shadow state.

🔒 IBM has disclosed a critical authentication bypass in IBM API Connect, tracked as CVE-2025-13915 with a CVSS score of 9.8. The flaw could allow remote attackers to gain unauthorized access to the application. Affected releases include 10.0.8.0–10.0.8.5 and 10.0.11.0. IBM advises downloading the interim fix from Fix Central and, if immediate patching is not possible, disabling Developer Portal self-service sign-up as a temporary mitigation.

🔒 IBM urged customers to patch a critical authentication bypass in its API Connect platform that could allow attackers to access applications remotely. Tracked as CVE-2025-13915 and rated 9.8/10, the flaw affects versions 10.0.11.0 and 10.0.8.0–10.0.8.5. Exploitation is low-complexity and requires no user interaction. IBM recommends upgrading to the latest release and offers interim mitigations, including disabling self-service sign-up on the Developer Portal.