vLLM is an inference and serving engine for large language models (LLMs). In versions starting from 0.7.0 to before 0.9.0, in the file vllm/multimodal/hasher.py, the MultiModalHasher class has a security and data integrity issue in its image hashing method. Currently, it serializes PIL.Image.Image objects using only obj.tobytes(), which returns only the raw pixel data, without including metadata such as the image’s shape (width, height, mode). As a result, two images of different sizes (e.g., 30x100 and 100x30) with the same pixel byte sequence could generate the same hash value. This may lead to hash collisions, incorrect cache hits, and even data leakage or security risks. This issue has been patched in version 0.9.0.

A denial of service vulnerability exists in the g_assert_not_reached functionality of Entr'ouvert Lasso 2.5.1 and 2.8.2. A specially crafted SAML assertion response can lead to a denial of service. An attacker can send a malformed SAML response to trigger this vulnerability.

Improper Handling of Case Sensitivity vulnerability in Apache Tomcat's GCI servlet allows security constraint bypass of security constraints that apply to the pathInfo component of a URI mapped to the CGI servlet. This issue affects Apache Tomcat: from 11.0.0-M1 through 11.0.6, from 10.1.0-M1 through 10.1.40, from 9.0.0.M1 through 9.0.104. The following versions were EOL at the time the CVE was created but are known to be affected: 8.5.0 though 8.5.100. Other, older, EOL versions may also be affected. Users are recommended to upgrade to version 11.0.7, 10.1.41 or 9.0.105, which fixes the issue.

Formidable (aka node-formidable) 2.1.0 through 3.x before 3.5.3 relies on hexoid to prevent guessing of filenames for untrusted executable content; however, hexoid is documented as not "cryptographically secure." (Also, there is a scenario in which only the last two characters of a hexoid string need to be guessed, but this is not often relevant.) NOTE: this does not imply that, in a typical use case, attackers will be able to exploit any hexoid behavior to upload and execute their own content.

In Artifex Ghostscript before 10.05.0, decode_utf8 in base/gp_utf8.c mishandles overlong UTF-8 encoding. NOTE: this issue exists because of an incomplete fix for CVE-2024-46954.

OpenPLC 3 through 64f9c11 has server.cpp Memory Corruption because a thread may access handleConnections arguments after the parent stack frame becomes unavailable.

vLLM is an inference and serving engine for large language models (LLMs). Prior to version 0.9.0, when a new prompt is processed, if the PageAttention mechanism finds a matching prefix chunk, the prefill process speeds up, which is reflected in the TTFT (Time to First Token). These timing differences caused by matching chunks are significant enough to be recognized and exploited. This issue has been patched in version 0.9.0.

Vault Community and Vault Enterprise rekey and recovery key operations can lead to a denial of service due to uncontrolled cancellation by a Vault operator. This vulnerability (CVE-2025-4656) has been remediated in Vault Community Edition 1.20.0 and Vault Enterprise 1.20.0, 1.19.6, 1.18.11, 1.17.17, and 1.16.22.

Open Policy Agent (OPA) is an open source, general-purpose policy engine. Prior to version 1.4.0, when run as a server, OPA exposes an HTTP Data API for reading and writing documents. Requesting a virtual document through the Data API entails policy evaluation, where a Rego query containing a single data document reference is constructed from the requested path. This query is then used for policy evaluation. A HTTP request path can be crafted in a way that injects Rego code into the constructed query. The evaluation result cannot be made to return any other data than what is generated by the requested path, but this path can be misdirected, and the injected Rego code can be crafted to make the query succeed or fail; opening up for oracle attacks or, given the right circumstances, erroneous policy decision results. Furthermore, the injected code can be crafted to be computationally expensive, resulting in a Denial Of Service (DoS) attack. This issue has been patched in version 1.4.0....

Vite is a frontend tooling framework for javascript. Prior to versions 6.3.4, 6.2.7, 6.1.6, 5.4.19, and 4.5.14, the contents of files in the project root that are denied by a file matching pattern can be returned to the browser. Only apps explicitly exposing the Vite dev server to the network (using --host or server.host config option) are affected. Only files that are under project root and are denied by a file matching pattern can be bypassed. `server.fs.deny` can contain patterns matching against files (by default it includes .env, .env.*, *.{crt,pem} as such patterns). These patterns were able to bypass for files under `root` by using a combination of slash and dot (/.). This issue has been patched in versions 6.3.4, 6.2.7, 6.1.6, 5.4.19, and 4.5.14.

vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Versions starting from 0.8.0 and prior to 0.8.5 are affected by a critical performance vulnerability in the input preprocessing logic of the multimodal tokenizer. The code dynamically replaces placeholder tokens (e.g., <|audio_|>, <|image_|>) with repeated tokens based on precomputed lengths. Due to ​​inefficient list concatenation operations​​, the algorithm exhibits ​​quadratic time complexity (O(n²))​​, allowing malicious actors to trigger resource exhaustion via specially crafted inputs. This issue has been patched in version 0.8.5.

JRuby-OpenSSL is an add-on gem for JRuby that emulates the Ruby OpenSSL native library. Starting in JRuby-OpenSSL version 0.12.1 and prior to version 0.15.4 (corresponding to JRuby versions starting in 9.3.4.0 prior to 9.4.12.1 and 10.0.0.0 prior to 10.0.0.1), when verifying SSL certificates, JRuby-OpenSSL does not verify that the hostname presented in the certificate matches the one the user tries to connect to. This means a man-in-the-middle could just present any valid cert for a completely different domain they own, and JRuby would accept the cert. Anybody using JRuby to make requests of external APIs, or scraping the web, that depends on https to connect securely. JRuby-OpenSSL version 0.15.4 contains a fix for the issue. This fix is included in JRuby versions 10.0.0.1 and 9.4.12.1.

A flaw was found in libsoup. When libsoup clients encounter an HTTP redirect, they mistakenly send the HTTP Authorization header to the new host that the redirection points to. This allows the new host to impersonate the user to the original host that issued the redirect.

A flaw was found in libsoup. It is vulnerable to memory leaks in the soup_header_parse_quality_list() function when parsing a quality list that contains elements with all zeroes.

A denial of service vulnerability exists in the lasso_provider_verify_saml_signature functionality of Entr&#39;ouvert Lasso 2.5.1. A specially crafted SAML response can lead to a denial of service. An attacker can send a malformed SAML response to trigger this vulnerability.

In xfig diagramming tool, a segmentation fault while running fig2dev allows an attacker to availability via local input manipulation via read_arcobject function.

A flaw was found in fig2dev. This vulnerability allows availability via local input manipulation via genge_itp_spline function.

In xfig diagramming tool, a stack-overflow while running fig2dev allows memory corruption via local input manipulation via read_objects function.

A flaw was found in xfig. This vulnerability allows possible code execution via local input manipulation via bezier_spline function.

In multispectral MIFF image processing in ImageMagick before 7.1.1-44, packet_size is mishandled (related to the rendering of all channels in an arbitrary order).