CVE Feed

A security flaw has been discovered in bigsk1 openai-realtime-ui up to 188ccde27fdf3d8fab8da81f3893468f53b2797c. The affected element is an unknown function of the file server.js of the component API Proxy Endpoint. Performing a manipulation of the argument Query results in server-side request forgery. The attack can be initiated remotely. The exploit has been released to the public and may be used for attacks. Continious delivery with rolling releases is used by this product. Therefore, no version details of affected nor updated releases are available. The patch is named 54f8f50f43af97c334a881af7b021e84b5b8310f. It is suggested to install a patch to address this issue.

NiceGUI is a Python-based UI framework. Prior to 3.10.0, Since PurePosixPath only recognizes forward slashes (/) as path separators, an attacker can bypass this sanitization on Windows by using backslashes (\) in the upload filename. Applications that construct file paths using file.name (a pattern demonstrated in NiceGUI's bundled examples) are vulnerable to arbitrary file write on Windows. This vulnerability is fixed in 3.10.0.

AIL framework is an open-source platform to collect, crawl, process and analyse unstructured data. Prior to 6.8, a stored cross-site scripting (XSS) vulnerability was identified in the modal item preview functionality. When item content longer than 800 characters was processed, attacker-controlled content was returned without an explicit text/plain content type, allowing the browser to interpret the response as active HTML. This could result in execution of arbitrary JavaScript in the context of an authenticated user viewing a crafted item. This vulnerability is fixed in 6.8.

Frappe Learning Management System (LMS) is a learning system that helps users structure their content. Prior to 2.46.0, a vulnerability has been identified in Frappe Learning where quiz scores can be modified by students before submission. The application currently relies on client-side calculated scores, which can be altered using browser developer tools prior to sending the submission request. While this does not allow modification of other users’ data or privilege escalation, it compromises the integrity of quiz results and undermines academic reliability. This issue affects data integrity but does not expose confidential information or allow unauthorized access to other accounts. This vulnerability is fixed in 2.46.0.

MinIO is a high-performance object storage system. From RELEASE.2018-08-18T03-49-57Z to before RELEASE.2025-12-20T04-58-37Z, MinIO's S3 Select feature is vulnerable to memory exhaustion when processing CSV files containing lines longer than available memory. The CSV reader's nextSplit() function calls bufio.Reader.ReadBytes('\n') with no size limit, buffering the entire input in memory until a newline is found. A CSV file with no newline characters causes the entire contents to be read into a single allocation, leading to an OOM crash of the MinIO server process. This is exploitable by any authenticated user with s3:PutObject and s3:GetObject permissions. The attack is especially practical when combined with compression: a ~2 MB gzip-compressed CSV can decompress to gigabytes of data without newlines, allowing a small upload to cause large memory consumption on the server. However, compression is not required — a sufficiently large uncompressed CSV with no newlines triggers the same issue.

Remnawave Backend is the backend for the Remnawave proxy and user management solution. Prior to 2.7.5, a glitch in the HWID device registration logic allows an authenticated user to bypass the configured limit for HWID devices and register more devices than expected, allowing them to resell subscriptions and consume excessive traffic. This vulnerability is fixed in 2.7.5.

A vulnerability was identified in idachev mcp-javadc up to 1.2.4. Impacted is an unknown function of the component HTTP Interface. Such manipulation of the argument jarFilePath leads to os command injection. It is possible to launch the attack remotely. The exploit is publicly available and might be used. The project was informed of the problem early through an issue report but has not responded yet.

Kamailio is an open source implementation of a SIP Signaling Server. Prior to 6.0.5 and 5.8.7, an out-of-bounds read in the auth module of Kamailio (formerly OpenSER and SER) allows remote attackers to cause a denial of service (process crash) via a specially crafted SIP packet if a successful user authentication without a database backend is followed by additional user identity checks. This vulnerability is fixed in 6.0.5 and 5.8.7.

Kamailio is an open source implementation of a SIP Signaling Server. Prior to 6.1.1, 6.0.6, and 5.8.8, an out-of-bounds access in the core of Kamailio (formerly OpenSER and SER) allows remote attackers to cause a denial of service (process crash) via a specially crafted data packet sent over TCP. The issue impacts Kamailio instances having TCP or TLS listeners. This vulnerability is fixed in 5.1.1, 6.0.6, and 5.8.8.

Tophat is a mobile applications testing harness. Prior to 2.5.1, Tophat is affected by remote code execution via crafted tophat:// or http://localhost:29070 URLs. The arguments query parameter flows unsanitized from URL parsing through to /bin/bash -c execution, allowing an attacker to execute arbitrary commands on a developer's macOS workstation. Any developer with Tophat installed is vulnerable. For previously trusted build hosts, no confirmation dialog appears. Attacker commands run with the user's permissions. This vulnerability is fixed in 2.5.1.

LiquidJS is a Shopify / GitHub Pages compatible template engine in pure JavaScript. Prior to 10.25.3, liquidjs 10.25.0 documents root as constraining filenames passed to renderFile() and parseFile(), but top-level file loads do not enforce that boundary. A Liquid instance configured with an empty temporary directory as root can return the contents of arbitrary files. This vulnerability is fixed in 10.25.3.

LightRAG provides simple and fast retrieval-augmented generation. Prior to 1.4.14, the LightRAG API is vulnerable to a JWT algorithm confusion attack where an attacker can forge tokens by specifying 'alg': 'none' in the JWT header. Since the jwt.decode() call does not explicitly deny the 'none' algorithm, a crafted token without a signature will be accepted as valid, leading to unauthorized access. This vulnerability is fixed in 1.4.14.

LiquidJS is a Shopify / GitHub Pages compatible template engine in pure JavaScript. Prior to 10.25.4, the sort_natural filter bypasses the ownPropertyOnly security option, allowing template authors to extract values of prototype-inherited properties through a sorting side-channel attack. Applications relying on ownPropertyOnly: true as a security boundary (e.g., multi-tenant template systems) are exposed to information disclosure of sensitive prototype properties such as API keys and tokens. This vulnerability is fixed in 10.25.4.

LobeHub is a work-and-lifestyle space to find, build, and collaborate with agent teammates that grow with you. Prior to 2.1.48, the webapi authentication layer trusts a client-controlled X-lobe-chat-auth header that is only XOR-obfuscated, not signed or otherwise authenticated. Because the XOR key is hardcoded in the repository, an attacker can forge arbitrary auth payloads and bypass authentication on protected webapi routes. Affected routes include /webapi/chat/[provider], /webapi/models/[provider], /webapi/models/[provider]/pull, and /webapi/create-image/comfyui. This vulnerability is fixed in 2.1.48.

InvenTree is an Open Source Inventory Management System. Prior to 1.2.7 and 1.3.0, when INVENTREE_DOWNLOAD_FROM_URL is enabled (opt-in), authenticated users can supply remote_image URLs that are fetched server-side via requests.get() with only Django's URLValidator check. There is no validation against private IP ranges or internal hostnames. Redirects are followed (allow_redirects=True), enabling bypass of any URL-format checks. This vulnerability is fixed in 1.2.7 and 1.3.0.

LiquidJS is a Shopify / GitHub Pages compatible template engine in pure JavaScript. Prior to 10.25.3, for {% include %}, {% render %}, and {% layout %}, LiquidJS checks whether the candidate path is inside the configured partials or layouts roots before reading it. That check is path-based, not realpath-based. Because of that, a file like partials/link.liquid passes the directory containment check as long as its pathname is under the allowed root. If link.liquid is actually a symlink to a file outside the allowed root, the filesystem follows the symlink when the file is opened and LiquidJS renders the external target. So the restriction is applied to the path string that was requested, not to the file that is actually read. This matters in environments where an attacker can place templates or otherwise influence files under a trusted template root, including uploaded themes, extracted archives, mounted content, or repository-controlled template trees. This vulnerability is fixed in 10.25.3.

InvenTree is an Open Source Inventory Management System. Prior to 1.2.7 and 1.3.0, any users who have staff access permissions can install plugins via the API, without requiring "superuser" account access. This level of permission requirement is out of alignment with other plugin actions (such as uninstalling) which do require superuser access. The vulnerability allows staff users (who may be considered to have a lower level of trust than a superuser account) to install arbitrary (and potentially harmful) plugins. This vulnerability is fixed in 1.2.7 and 1.3.0.

InvenTree is an Open Source Inventory Management System. Prior to 1.2.7 and 1.3.0, a non-staff authenticated user can elevate their account to a staff level via a POST request against their user account endpoint. The write permissions on the API endpoint are improperly configured, allowing any user to change their staff status. This vulnerability is fixed in 1.2.7 and 1.3.0.

InvenTree is an Open Source Inventory Management System. From 0.16.0 to before 1.2.7, any authenticated InvenTree user can create a valid API token attributed to any other user in the system — including administrators and superusers — by supplying the target's user ID in the user field of a POST /api/user/tokens/ request. The returned token is immediately usable for full API authentication as the target user, from any network location, with no further interaction required. This vulnerability is fixed in 1.2.7 and 1.3.0.

InvenTree is an Open Source Inventory Management System. From 1.2.3 to 1.2.6, the fix for CVE-2026-27629 upgraded the PART_NAME_FORMAT validator to use jinja2.sandbox.SandboxedEnvironment. However, the actual renderer in part/helpers.py was not updated and still uses the non-sandboxed jinja2.Environment. Additionally, the validator uses a dummy Part instance with pk=None, which allows conditional template expressions to behave differently during validation versus production rendering. A staff user with settings access can craft a template that passes validation but executes arbitrary code during rendering. This issue requires access by a user with granted staff permissions. This vulnerability is fixed in 1.2.7 and 1.3.0.

A denial of service vulnerability exists in React Server Components, affecting the following packages: react-server-dom-parcel, react-server-dom-turbopack and react-server-dom-webpack (versions 19.0.0 through 19.0.4, 19.1.0 through 19.1.5, and 19.2.0 through 19.2.4). The vulnerability is triggered by sending specially crafted HTTP requests to Server Function endpoints.The payload of the HTTP request causes excessive CPU usage for up to a minute ending in a thrown error that is catchable.

immich is a high performance self-hosted photo and video management solution. Prior to 2.7.0, sStored Cross-Site Scripting (XSS) in the 360° panorama viewer allows any authenticated user to execute arbitrary JavaScript in the browser of any other user who views the malicious panorama with the OCR overlay enabled. The attacker uploads an equirectangular image containing crafted text; OCR extracts it, and the panorama viewer renders it via innerHTML without sanitization. This enables session hijacking (via persistent API key creation), private photo exfiltration, and access to GPS location history and face biometric data. This vulnerability is fixed in 2.7.0.

LORIS (Longitudinal Online Research and Imaging System) is a self-hosted web application that provides data- and project-management for neuroimaging research. From 24.0.0 to before 27.0.3 and 28.0.1, an incorrect order of operations in the FilesDownloadHandler could result in an attacker escaping the intended download directories. This vulnerability is fixed in 27.0.3 and 28.0.1.

LORIS (Longitudinal Online Research and Imaging System) is a self-hosted web application that provides data- and project-management for neuroimaging research. From 15.10 to before 27.0.3 and 28.0.1, there is a potential for a cross-site scripting attack in the survey_accounts module if a user provides an invalid visit label. While the data is properly JSON encoded, the Content-Type header is not set causing the web browser to interpret the payload as HTML, opening the possibility of a cross-site scripting if a user is tricked into following an invalid link. This vulnerability is fixed in 27.0.3 and 28.0.1.

LORIS (Longitudinal Online Research and Imaging System) is a self-hosted web application that provides data- and project-management for neuroimaging research. From 20.0.0 to before 27.0.3 and 28.0.1, an endpoint in the publication module was incorrectly trusting the baseURL submitted by a user's POST request rather than the internal LORIS value. This could result in a theoretical attacker with publication module access forging an email to an external domain under the attacker's control which appeared to come from LORIS. This vulnerability is fixed in 27.0.3 and 28.0.1.

LORIS (Longitudinal Online Research and Imaging System) is a self-hosted web application that provides data- and project-management for neuroimaging research. From to before 27.0.3 and 28.0.1, the help_editor module of LORIS did not properly sanitize some user supplied variables which could result in a reflected cross-site scripting attack if a user is tricked into following an invalid link. The same input vector could also allow an attacker to download arbitrary markdown files on an unpatched server. This vulnerability is fixed in 27.0.3 and 28.0.1.

LORIS (Longitudinal Online Research and Imaging System) is a self-hosted web application that provides data- and project-management for neuroimaging research. From 21.0.0 to before 27.0.3 and 28.0.1, while the document_repository frontend was restricting file access, the backend endpoint was not correctly verifying access permissions. A user could theoretically download a file that they should not have access to, if they know or can brute force the filename. This vulnerability is fixed in 27.0.3 and 28.0.1.

LORIS (Longitudinal Online Research and Imaging System) is a self-hosted web application that provides data- and project-management for neuroimaging research. From 16.1.0 to before 27.0.3 and 28.0.1, While the frontend of the media module filters files that the user should not have access to, the backend was not applying access checks and it would be possible for someone who should not have access to a file to access it if they know the filename. This vulnerability is fixed in 27.0.3 and 28.0.1.

Use of Default Cryptographic Key in the hardware for some Intel(R) Pentium(R) Processor Silver Series, Intel(R) Celeron(R) Processor J Series, Intel(R) Celeron(R) Processor N Series may allow an escalation of privilege. Hardware reverse engineer adversary with a privileged user combined with a high complexity attack may enable escalation of privilege. This result may potentially occur via physical access when attack requirements are present with special internal knowledge and requires no user interaction. The potential vulnerability may impact the confidentiality (high), integrity (none) and availability (none) of the vulnerable system, resulting in subsequent system confidentiality (high), integrity (high) and availability (none) impacts.

Zammad is a web based open source helpdesk/customer support system. Prior to 7.0.1, he REST endpoint POST /api/v1/ai_assistance/text_tools/:id contains an authorization failure. Context data (e.g., a group or organization) supplied to be used in the AI prompt were not checked if they are accessible for the current user. This leads to having data present in the AI prompt that were not authorized before being used. A user needs to have ticket.agent permission to be able to use the provided context data. This vulnerability is fixed in 7.0.1.

Zammad is a web based open source helpdesk/customer support system. Prior to 7.0.1 and 6.5.4, the REST endpoint POST /api/v1/ai_assistance/text_tools/:id was not checking if a user is privileged to use the text tool, resulting in being able to use it in all situations. This vulnerability is fixed in 7.0.1 and 6.5.4.

Zammad is a web based open source helpdesk/customer support system. Prior to 7.0.1, a server-side template injection vulnerability which leads to RCE via AI Agent exists. Impact is limited to environments where an attacker can control or influence type_enrichment_data (typically high-privilege administrative configuration). This vulnerability is fixed in 7.0.1.

Zammad is a web based open source helpdesk/customer support system. Prior to 7.0.1 and 6.5.4, unauthenticated remote attackers were able to access the getting started endpoint to get access to sensitive internal entity data, even after the system setup was completed. This vulnerability is fixed in 7.0.1 and 6.5.4.

Zammad is a web based open source helpdesk/customer support system. Prior to 7.0.1 and 6.5.4, the used endpoint for ticket creation was missing authorization if the related parameter for adding links is used. This vulnerability is fixed in 7.0.1 and 6.5.4.

Zammad is a web based open source helpdesk/customer support system. Prior to 7.0.1 and 6.5.4, the OAuth callback endpoints for Microsoft, Google, and Facebook external credentials do not validate a CSRF state parameter. This vulnerability is fixed in 7.0.1 and 6.5.4.

Zammad is a web based open source helpdesk/customer support system. Prior to 7.0.1 and 6.5.4, the SSO mechanism in Zammad was not verifying the header originates from a trusted SSO proxy/gateway before applying further actions on it. This vulnerability is fixed in 7.0.1 and 6.5.4.

Zammad is a web based open source helpdesk/customer support system. Prior to 7.0.1 and 6.5.4, the webhook model was missing a proper validation for loop back addresses, or link-local addresses — only the URL scheme (HTTP/HTTPS) as well as the hostname was checked. This could end up in retrieving confidential metadata of cloud/hosting providers. The existing check is now extended and is applied when configuring webhooks as well as triggering webhook jobs. This vulnerability is fixed in 7.0.1 and 6.5.4.

Zammad is a web based open source helpdesk/customer support system. Prior to 7.0.1 and 6.5.4, the HTML sanitizer for ticket articles was missing proper sanitization of data: ... URI schemes, resulting in storing such malicious content in the database of the Zammad instance. The Zammad GUI is rendering this content, due to applied CSP rules no harm was done by e.g., clicking such a link. This vulnerability is fixed in 7.0.1 and 6.5.4.

Zammad is a web based open source helpdesk/customer support system. Prior to 7.0.1, customers in shared organizations (means they can see each other's tickets) could see fields which are not intended for customers - including fields not intended for them at all (e.g. priority, custom ticket attributes for internal purposes). This was the case when a customer opened a ticket from another user of the same shared organization. They are not able to modify these field. This vulnerability is fixed in 7.0.1.

LORIS (Longitudinal Online Research and Imaging System) is a self-hosted web application that provides data- and project-management for neuroimaging research. From 20.0.0 to before 27.0.3 and 28.0.1, a bug in the static file router can allow an attacker to traverse outside of the intended directory, allowing unintended files to be downloaded through the static, css, and js endpoints. This vulnerability is fixed in 27.0.3 and 28.0.1.

An OS command injection vulnerability in the dnsmasq module of TP-Link Archer AX53 v1.0 allows an authenticated adjacent attacker to execute arbitrary code when a specially crafted configuration file is processed due to insufficient input validation. Successful exploitation may allow the attacker to modify device configuration, access sensitive information, or further compromise system integrity. This issue affects AX53 v1.0: before 1.7.1 Build 20260213.

An external configuration control vulnerability in the OpenVPN module of TP-Link AX53 v1.0 allows an authenticated adjacent attacker to read arbitrary files when a malicious configuration file is processed. Successful exploitation may allow unauthorized access to arbitrary files on the device, potentially exposing sensitive information.This issue affects AX53 v1.0: before 1.7.1 Build 20260213.

An external control of configuration vulnerability in the OpenVPN module of TP-Link AX53 v1.0 allows an authenticated adjacent attacker to read arbitrary file when a malicious configuration file is processed.  Successful exploitation may allow unauthorized access to arbitrary files on the device, potentially exposing sensitive information.This issue affects AX53 v1.0: before 1.7.1 Build 20260213.

An OS command injection vulnerability in the OpenVPN module of TP-Link Archer AX53 v1.0 allows an authenticated adjacent attacker to execute system commands when a specially crafted configuration file is processed due to insufficient input validation. Successful exploitation may allow modification of configuration files, disclosure of sensitive information, or further compromise of device integrity. This issue affects AX53 v1.0: before 1.7.1 Build 20260213.

LiquidJS is a Shopify / GitHub Pages compatible template engine in pure JavaScript. Prior to 10.25.3, the replace filter in LiquidJS incorrectly accounts for memory usage when the memoryLimit option is enabled. It charges str.length + pattern.length + replacement.length bytes to the memory limiter, but the actual output from str.split(pattern).join(replacement) can be quadratically larger when the pattern occurs many times in the input string. This allows an attacker who controls template content to bypass the memoryLimit DoS protection with approximately 2,500x amplification, potentially causing out-of-memory conditions. This vulnerability is fixed in 10.25.3.

A stack-based buffer overflow in the tmpServer module of TP-Link Archer AX53 v1.0 allows an authenticated adjacent attacker to trigger a segmentation fault and potentially execute arbitrary code via a specially crafted configuration file. Successful exploitation may cause a crash and could allow arbitrary code execution, enabling modification of device state, exposure of sensitive data, or further compromise of device integrity. This issue affects AX53 v1.0: before 1.7.1 Build 20260213.

LORIS (Longitudinal Online Research and Imaging System) is a self-hosted web application that provides data- and project-management for neuroimaging research. Prior to 27.0.3 and 28.0.1, a SQL injection has been identified in some code sections for the MRI feedback popup window of the imaging browser. Attackers can use SQL ingestion to access/alter data on the server. This vulnerability is fixed in 27.0.3 and 28.0.1.

Fleet is open source device management software. Prior to 4.81.1, the Orbit agent's FileVault disk encryption key rotation flow on collects a local user's password via a GUI dialog and interpolates it directly into a Tcl/expect script executed via exec.Command("expect", "-c", script). Because the password is inserted into Tcl brace-quoted send {%s}, a password containing } terminates the literal and injects arbitrary Tcl commands. Since Orbit runs as root, this allows a local unprivileged user to escalate to root privileges. This vulnerability is fixed in 4.81.1.

Saleor is an e-commerce platform. From 2.10.0 to before 3.23.0a3, 3.22.47, 3.21.54, and 3.20.118, the requestEmailChange() mutation was revealing the existence of user-provided email addresses in error messages. This vulnerability is fixed in 3.23.0a3, 3.22.47, 3.21.54, and 3.20.118.

A Missing Authentication for Critical Function vulnerability in command processing of Juniper Networks Junos OS allows a privileged local attacker to gain access to Linux-based line cards as root. This issue affects systems running Junos OS using Linux-based line cards. Affected line cards include: * MPC7, MPC8, MPC9, MPC10, MPC11 * LC2101, LC2103 * LC480, LC4800, LC9600 * MX304 (built-in FPC) * MX-SPC3 * SRX5K-SPC3 * EX9200-40XS * FPC3-PTX-U2, FPC3-PTX-U3 * FPC3-SFF-PTX * LC1101, LC1102, LC1104, LC1105 This issue affects Junos OS:  * all versions before 22.4R3-S8,  * from 23.2 before 23.2R2-S6,  * from 23.4 before 23.4R2-S6,  * from 24.2 before 24.2R2-S3,  * from 24.4 before 24.4R2, * from 25.2 before 25.2R2.