CVE-2026-91139 is a cryptographic vulnerability in TP-Link Systems Inc. Deco BE65 v3 firmware. The local web-management authentication flow uses an insufficient 512-bit RSA key, which weakens the cryptographic protection. This weakness could allow attackers to compromise the authentication traffic. The issue affects version 0 of the firmware and is fixed in version 1.0.11 Build 26092212 Rel. 18351. The CVSS 4.0 score is 6.1, indicating a medium severity level.
django-filer versions up to 3.4.4 with per-folder permissions enabled (FILER_ENABLE_PERMISSIONS=True) have a vulnerability where the admin 'Delete files and/or folders' action does not enforce per-folder edit permissions. This allows staff users with global model-level delete permissions but read-only folder-scoped permissions to delete files and folders unauthorized. The issue is fixed in version 3.5.0.
djangocms-versioning versions prior to 2.6.1 contain an authorization bypass vulnerability where two version state transitions, discard and edit_redirect, do not enforce per-object change permissions. This allows any staff user to discard (delete) draft versions and their underlying content objects or create editable drafts from published versions without proper authorization. The issue arises because these transitions lack the user_can_change condition that other transitions enforce, resulting in broken access control. The vulnerability affects multi-editor deployments with differentiated per-object permissions, reducing authorization to is_staff or a weaker global change permission. The flaw is fixed in version 2.6.1 by adding the missing user_can_change condition to these transitions.
LLMKube versions prior to 0.9.0 contain a vulnerability in the ModelReconciler component where the Model.spec.source field accepts arbitrary absolute paths and file:// URIs without path allowlisting. This allows users with create/update permissions on Model and InferenceService custom resources to read arbitrary files accessible to the controller pod or on the node hosting the inference pod, including sensitive files like Kubernetes service-account tokens and kubelet configuration. Model caching, which triggers the vulnerable hostPath volume mount, is enabled by default. The vulnerability has a CVSS 3.1 base score of 7.7 (High).
The install-artifact-from-github package versions prior to 1.7.0 download native addon binaries over the network without any cryptographic integrity verification. The download mirror URL can be fully controlled via an environment variable, allowing an attacker to redirect the download to a malicious host. The downloaded file is written directly to disk and then loaded as a native module, enabling arbitrary code execution if the artifact is tampered with. Additionally, the package allows plaintext HTTP downloads and silent downgrades from HTTPS to HTTP via redirects, increasing the risk of man-in-the-middle attacks. Post-download verification steps do not prevent malicious code execution as they run after the binary is loaded. A fix is available in version 1.7.0 that addresses these issues.
A vulnerability in NGINX Plus and NGINX Open Source occurs when configured with try_files and alias directives using a map variable, causing certain regex captures to trigger a segmentation fault in a worker process. This results in the unexpected termination of the NGINX worker process. The issue affects specific Ubuntu package versions of NGINX. There is no control plane exposure; the impact is limited to the data plane.
A vulnerability in NGINX Plus and NGINX Open Source configured with the ngx_stream_ssl_preread_module allows an unauthenticated attacker to send a specially crafted ClientHello message that forges module variable values. This can affect virtual server selection and potentially route connections to privileged or internal backends that should be inaccessible, exposing sensitive information. This issue affects data plane routing only and does not expose the control plane. The vulnerability has a medium severity with a CVSS score of 5.3.
A use-after-free vulnerability exists in NGINX Plus and NGINX Open Source when configured as a proxy with a proxied server name containing variables, combined with Server Side Includes (SSI) or mirror directives that cause subrequests resolving the same server name. This flaw may allow an attacker to crash the worker process, corrupt memory, or potentially execute code if ASLR is disabled or bypassed. The vulnerability affects specific Ubuntu package versions of NGINX. It is rated high severity with a CVSS score of 8.1.
A vulnerability in NGINX Plus and NGINX Open Source occurs when HTTP/2 is enabled for client connections and requests are proxied to a backend. A specially crafted request can trigger a use-after-free condition in the worker process if it is running low on worker_connections. This flaw can lead to denial-of-service or remote code execution on systems without ASLR or where ASLR is bypassed. The issue affects specific Ubuntu package versions of NGINX and does not expose the control plane, only the data plane.
A vulnerability in NGINX Plus and NGINX Open Source HTTP/2 configurations can cause the flow control window to exceed allowed limits, leading to a heap buffer overflow and potential worker process crash or remote code execution. This issue affects specific Ubuntu package versions of NGINX. The vulnerability requires conditions such as ASLR being disabled or bypassed for code execution. It is a data plane issue without control plane exposure.
CVE-2026-121810 affects NGINX Plus and NGINX Open Source configurations using HTTP/3. It allows a remote unauthenticated attacker to cause a use-after-free condition in the worker process, potentially leading to a denial-of-service (DoS) or code execution on systems with ASLR disabled or bypassed. This vulnerability impacts only the data plane and does not expose the control plane. The issue affects specific Ubuntu package versions of NGINX. The CVSS score is 8.1, indicating high severity.
A vulnerability in NGINX Plus and NGINX Open Source occurs when configured to require client certificates (ssl_verify_client on) in a non-default stream virtual server. This flaw allows a remote attacker to bypass client certificate verification, impacting the data plane only without exposing the control plane. The issue affects specific Ubuntu package versions of NGINX. The severity is rated medium with a CVSS score of 4.8.
A vulnerability in NGINX occurs when the listen directive is used with the ssl parameter in the http block with TLSv1.3 enabled. An attacker can send a specially crafted TLS connection that causes an infinite loop in the worker process while reading the HTTP request body, leading to denial-of-service (DoS) by stopping the worker from accepting or processing connections. This affects specific Ubuntu package versions of NGINX. There is no control plane exposure, only a data plane impact.
CVE-2026-113425 is a use-after-free vulnerability in NGINX Plus and NGINX Open Source when configured as a mail proxy. This flaw allows remote unauthenticated attackers to crash nginx worker processes or corrupt memory, potentially leading to remote code execution on systems without effective ASLR. The issue affects specific Ubuntu package versions of nginx and does not expose the control plane, only the data plane.
A vulnerability in DOMPurify when configured with SAFE_FOR_TEMPLATES and a DOM-returning mode allows template expressions split across text nodes in attached shadow roots to survive sanitization. This occurs because the post-walk safety net does not recurse into shadow roots, enabling attackers who supply DOM nodes with shadow roots to inject expressions that are later evaluated by template engines. Exploitation requires specific non-default configurations and attacker control over DOM nodes with attached shadow roots. The vulnerability is fixed in version 3.4.17.
DOMPurify versions prior to 3.4.17 have a vulnerability in the IN_PLACE sanitization mode where hooks that remove a parent or sibling node leave detached subtrees with active event handlers unneutralized. This allows event handlers such as <img onerror> to fire after sanitize() returns, potentially leading to DOM-based XSS. The issue arises because the sanitization process only neutralizes the node directly removed by the hook, not other detached nodes like parents or siblings. This behavior breaks the expected guarantee that detached subtrees are neutralized inline in IN_PLACE mode.
open-webui version 0.6.33 and earlier versions prior to 0.9.0 have an incorrect access control vulnerability. The API endpoint /api/tasks/stop/ allows any authenticated user to stop arbitrary language model response tasks without verifying task ownership. This flaw enables normal users to interfere with other users' tasks.
A security update for glib2 addresses a buffer under-read vulnerability identified as CVE-2025-7039. The flaw exists in the glib/gfileutils.c component, specifically in the get_tmp_file() function. This issue could potentially lead to reading memory before the buffer, which may cause unexpected behavior or information disclosure. The vulnerability is rated with medium severity. No CVSS score is provided, and there is no indication of exploitation in the wild. The update fixes this issue in affected versions of the SUSE product.
CVE-2025-51736 is a file upload vulnerability affecting HCL Technologies Ltd. Unica version 12.0.0. The vulnerability allows an attacker with low privileges to upload files that could lead to limited confidentiality, integrity, and availability impacts. The vulnerability has a medium severity rating with a CVSS score of 6.3.
An Improper Privilege Management vulnerability exists in Apache Kvrocks versions 2.9.0 through 2.13.0. This flaw allows users with limited privileges to perform actions beyond their authorization level. The issue is resolved in version 2.14.0, which users are advised to upgrade to.