vLLM is an inference and serving engine for large language models (LLMs). In versions 0.8.0 up to but excluding 0.9.0, hitting the /v1/completions API with a invalid json_schema as a Guided Param kills the vllm server. This vulnerability is similar GHSA-9hcf-v7m4-6m2j/CVE-2025-48943, but for regex instead of a JSON schema. Version 0.9.0 fixes the issue.

vLLM, an inference and serving engine for large language models (LLMs), has a Regular Expression Denial of Service (ReDoS) vulnerability in the file `vllm/entrypoints/openai/tool_parsers/pythonic_tool_parser.py` of versions 0.6.4 up to but excluding 0.9.0. The root cause is the use of a highly complex and nested regular expression for tool call detection, which can be exploited by an attacker to cause severe performance degradation or make the service unavailable. The pattern contains multiple nested quantifiers, optional groups, and inner repetitions which make it vulnerable to catastrophic backtracking. Version 0.9.0 contains a patch for the issue.

ModSecurity is an open source, cross platform web application firewall (WAF) engine for Apache, IIS and Nginx. Versions prior to 2.9.10 contain a denial of service vulnerability similar to GHSA-859r-vvv8-rm8r/CVE-2025-47947. The `sanitiseArg` (and `sanitizeArg` - this is the same action but an alias) is vulnerable to adding an excessive number of arguments, thereby leading to denial of service. Version 2.9.10 fixes the issue. As a workaround, avoid using rules that contain the `sanitiseArg` (or `sanitizeArg`) action.

A flaw was found in GIMP when processing XCF image files. If a user opens one of these image files that has been specially crafted by an attacker, GIMP can be tricked into making serious memory errors, potentially leading to crashes and causing use-after-free issues.

A flaw was found in GIMP when processing certain TGA image files. If a user opens one of these image files that has been specially crafted by an attacker, GIMP can be tricked into making serious memory errors, potentially leading to crashes and causing a heap buffer overflow.

A flaw was found in GIMP. The GIMP ani_load_image() function is vulnerable to a stack-based overflow. If a user opens.ANI files, GIMP may be used to store more information than the capacity allows. This flaw allows a malicious ANI file to trigger arbitrary code execution.

Improper Access Control vulnerability in Apache Commons. A special BeanIntrospector class was added in version 1.9.2. This can be used to stop attackers from using the declared class property of Java enum objects to get access to the classloader. However this protection was not enabled by default. PropertyUtilsBean (and consequently BeanUtilsBean) now disallows declared class level property access by default. Releases 1.11.0 and 2.0.0-M2 address a potential security issue when accessing enum properties in an uncontrolled way. If an application using Commons BeanUtils passes property paths from an external source directly to the getProperty() method of PropertyUtilsBean, an attacker can access the enum’s class loader via the “declaredClass” property available on all Java “enum” objects. Accessing the enum’s “declaredClass” allows remote attackers to access the ClassLoader and execute arbitrary code. The same issue exists with PropertyUtilsBean.getNestedProperty(). Sta...

gs_lib_ctx_stash_sanitized_arg in base/gslibctx.c in Artifex Ghostscript before 10.05.1 lacks argument sanitization for the # case. A created PDF document includes its password in cleartext.

An issue was discovered in Django 5.2 before 5.2.3, 5.1 before 5.1.11, and 4.2 before 4.2.23. Internal HTTP response logging does not escape request.path, which allows remote attackers to potentially manipulate log output via crafted URLs. This may lead to log injection or forgery when logs are viewed in terminals or processed by external systems.

tar-fs provides filesystem bindings for tar-stream. Versions prior to 3.0.9, 2.1.3, and 1.16.5 have an issue where an extract can write outside the specified dir with a specific tarball. This has been patched in versions 3.0.9, 2.1.3, and 1.16.5. As a workaround, use the ignore option to ignore non files/directories.

Next.js is a React framework for building full-stack web applications. In versions starting from 13.0 to before 14.2.30 and 15.0.0 to before 15.2.2, Next.js may have allowed limited source code exposure when the dev server was running with the App Router enabled. The vulnerability only affects local development environments and requires the user to visit a malicious webpage while npm run dev is active. This issue has been patched in versions 14.2.30 and 15.2.2.

jq is a command-line JSON processor. In versions up to and including 1.7.1, a heap-buffer-overflow is present in function `jv_string_vfmt` in the jq_fuzz_execute harness from oss-fuzz. This crash happens on file jv.c, line 1456 `void* p = malloc(sz);`. As of time of publication, no patched versions are available.

Radashi is a TypeScript utility toolkit. Prior to version 12.5.1, the set function within the Radashi library is vulnerable to prototype pollution. If an attacker can control parts of the path argument to the set function, they could potentially modify the prototype of all objects in the JavaScript runtime, leading to unexpected behavior, denial of service, or even remote code execution in some specific scenarios. This issue has been patched in version 12.5.1. A workaround for this issue involves sanitizing the path argument provided to the set function to ensure that no part of the path string is __proto__, prototype, or constructor.

In DOMPurify through 3.2.5 before 6bc6d60, scripts/server.js does not ensure that a pathname is located under the current working directory. NOTE: the Supplier disputes the significance of this report because the "Uncontrolled data used in path expression" occurs "in a development helper script which starts a local web server if needed and must be manually started."

Untrusted LD_LIBRARY_PATH environment variable vulnerability in the GNU C Library version 2.27 to 2.38 allows attacker controlled loading of dynamically shared library in statically compiled setuid binaries that call dlopen (including internal dlopen calls after setlocale or calls to NSS functions such as getaddrinfo).

CoreDNS is a DNS server that chains plugins. In versions prior to 1.12.2, a Denial of Service (DoS) vulnerability exists in the CoreDNS DNS-over-QUIC (DoQ) server implementation. The server previously created a new goroutine for every incoming QUIC stream without imposing any limits on the number of concurrent streams or goroutines. A remote, unauthenticated attacker could open a large number of streams, leading to uncontrolled memory consumption and eventually causing an Out Of Memory (OOM) crash — especially in containerized or memory-constrained environments. The patch in version 1.12.2 introduces two key mitigation mechanisms: `max_streams`, which caps the number of concurrent QUIC streams per connection with a default value of `256`; and `worker_pool_size`, which Introduces a server-wide, bounded worker pool to process incoming streams with a default value of `1024`. This eliminates the 1:1 stream-to-goroutine model and ensures that CoreDNS remains resilient under high concur...

ModSecurity is an open source, cross platform web application firewall (WAF) engine for Apache, IIS and Nginx. Versions up to and including 2.9.8 are vulnerable to denial of service in one special case (in stable released versions): when the payload's content type is `application/json`, and there is at least one rule which does a `sanitiseMatchedBytes` action. A patch is available at pull request 3389 and expected to be part of version 2.9.9. No known workarounds are available.

Argo CD is a declarative, GitOps continuous delivery tool for Kubernetes. Prior to versions 2.13.8, 2.14.13, and 3.0.4, an attacker can perform arbitrary actions on behalf of the victim via the API. Due to the improper filtering of URL protocols in the repository page, an attacker can achieve cross-site scripting with permission to edit the repository. This issue has been patched in versions 2.13.8, 2.14.13, and 3.0.4.

Varnish Cache before 7.6.3 and 7.7 before 7.7.1, and Varnish Enterprise before 6.0.13r14, allow client-side desync via HTTP/1 requests, because the product incorrectly permits CRLF to be skipped to delimit chunk boundaries.

Asterisk is an open-source private branch exchange (PBX). Prior to versions 18.26.2, 20.14.1, 21.9.1, and 22.4.1 of Asterisk and versions 18.9-cert14 and 20.7-cert5 of certified-asterisk, SIP requests of the type MESSAGE (RFC 3428) authentication do not get proper alignment. An authenticated attacker can spoof any user identity to send spam messages to the user with their authorization token. Abuse of this security issue allows authenticated attackers to send fake chat messages can be spoofed to appear to come from trusted entities. Even administrators who follow Security best practices and Security Considerations can be impacted. Therefore, abuse can lead to spam and enable social engineering, phishing and similar attacks. Versions 18.26.2, 20.14.1, 21.9.1, and 22.4.1 of Asterisk and versions 18.9-cert14 and 20.7-cert5 of certified-asterisk fix the issue.