ghsa

A flaw was found in the vLLM library. A completions API request with an empty prompt will crash the vLLM API server, resulting in a denial of service.

A vulnerability was found in the ilab model serve component, where improper handling of the best_of parameter in the vllm JSON web API can lead to a Denial of Service (DoS). The API used for LLM-based sentence or chat completion accepts a best_of parameter to return the best completion from several options. When this parameter is set to a large value, the API does not handle timeouts or resource exhaustion properly, allowing an attacker to cause a DoS by consuming excessive system resources. This leads to the API becoming unresponsive, preventing legitimate users from accessing the service.

### Summary A potential Cross-Site Scripting (XSS) vulnerability has been identified in the `/wireui/button` endpoint, specifically through the `label` query parameter. Malicious actors could exploit this vulnerability by injecting JavaScript into the `label` parameter, leading to the execution of arbitrary code in the victim's browser. ### Details The `/wireui/button` endpoint dynamically renders button labels based on user-provided input via the `label` query parameter. Due to insufficient sanitization or escaping of this input, an attacker can inject malicious JavaScript. The following URL demonstrates the vulnerability: ``` https://wireui.dev/wireui/button?label=Cancel&1%25%7ds8dk0%3E%3Cscript%3Ealert(1)%3C/script%3Ez1qt3=1 ``` By crafting such a request, an attacker can inject arbitrary code that will be executed by the browser when the endpoint is accessed. ### Proof of Concept (PoC) To demonstrate the vulnerability, visit the following URL: ``` /wireui/button?label=<script>...

### Impact An authenticated user can mute alert rules from arbitrary organizations and projects given a known given rule ID. The user does not need to be a member of the organization or have permissions on the project. In our review, we have identified no instances where alerts have been muted by unauthorized parties. ### Patches A patch was issued to ensure authorization checks are properly scoped on requests to mute alert rules. Authenticated users who do not have the necessary permissions are no longer able to mute alerts. Sentry SaaS users do not need to take any action. [Self-Hosted Sentry](https://github.com/getsentry/self-hosted) users should upgrade to version **24.9.0** or higher. ### Affected Versions The rule mute feature was generally available as of 23.6.0 but users with early access may have had the feature as of 23.4.0. ### References - [Prevent muting alerts](https://github.com/getsentry/sentry/pull/77016)

### Impact An authenticated user may delete user issue alert notifications for arbitrary users given a known alert ID. ### Patches A patch was issued to ensure authorization checks are properly scoped on requests to delete user alert notifications. Sentry SaaS users do not need to take any action. [Self-Hosted Sentry](https://github.com/getsentry/self-hosted) users should upgrade to version **24.9.0** or higher. ### References - [Prevent muting user alerts](https://github.com/getsentry/sentry/pull/77093/)

The czim/file-handling package before 1.5.0 and 2.x before 2.3.0 (used with PHP Composer) does not properly validate URLs within makeFromUrl and makeFromAny, leading to SSRF, and to directory traversal for the reading of local files.

An issue was discovered in the powermail extension through 12.4.0 for TYPO3. It fails to validate the mail parameter of the createAction, resulting in Insecure Direct Object Reference (IDOR) in some configurations. An unauthenticated attacker can use this to display user-submitted data of all forms persisted by the extension. The fixed versions are 7.5.1, 8.5.1, 10.9.1, and 12.4.1.

### Impact It is possible to inject insert tags in canonical URLs which will be replaced when the page is rendered. ### Patches Update to Contao 4.13.49, 5.3.15 or 5.4.3. ### Workarounds Disable canonical tags in the settings of the website root page. ### References https://contao.org/en/security-advisories/insert-tag-injection-via-canonical-urls ### For more information If you have any questions or comments about this advisory, open an issue in [contao/contao](https://github.com/contao/contao/issues/new/choose).

### Impact Back end users can list files outside their file mounts or the document root in the FileSelector widget. ### Patches Update to Contao 4.13.49. ### Workarounds None. ### References https://contao.org/en/security-advisories/directory-traversal-in-the-fileselector-widget ### For more information If you have any questions or comments about this advisory, open an issue in [contao/contao](https://github.com/contao/contao/issues/new/choose). ### Credits Thanks to Jakob Steeg from usd AG for reporting this vulnerability.

### Impact Back end users with access to the file manager can upload malicious files and execute them on the server. ### Patches Update to Contao 4.13.49, 5.3.15 or 5.4.3. ### Workarounds Configure your web server so it does not execute PHP files and other scripts in the Contao file upload directory. ### References https://contao.org/en/security-advisories/remote-command-execution-through-file-uploads ### For more information If you have any questions or comments about this advisory, open an issue in [contao/contao](https://github.com/contao/contao/issues/new/choose). ### Credits Thanks to Jakob Steeg from usd AG for reporting this vulnerability.