A vulnerability in the ObsidianReader class of the run-llama/llama_index repository, specifically in version 0.12.27, allows for hardlink-based path traversal. This flaw permits attackers to bypass path restrictions and access sensitive system files, such as /etc/passwd, by exploiting hardlinks. The vulnerability arises from inadequate handling of hardlinks in the load_data() method, where the security checks fail to differentiate between real files and hardlinks. This issue is resolved in version 0.5.2.

A path traversal vulnerability exists in run-llama/llama_index versions 0.12.27 through 0.12.40, specifically within the `encode_image` function in `generic_utils.py`. This vulnerability allows an attacker to manipulate the `image_path` input to read arbitrary files on the server, including sensitive system files. The issue arises due to improper validation or sanitization of the file path, enabling path traversal sequences to access files outside the intended directory. The vulnerability is fixed in version 0.12.41.

The `SimpleDirectoryReader` component in `llama_index.core` version 0.12.23 suffers from uncontrolled memory consumption due to a resource management flaw. The vulnerability arises because the user-specified file limit (`num_files_limit`) is applied after all files in a directory are loaded into memory. This can lead to memory exhaustion and degraded performance, particularly in environments with limited resources. The issue is resolved in version 0.12.41.

A malicious user may submit a specially-crafted complex payload that otherwise meets the default request size limit which results in excessive memory and CPU consumption of Vault. This may lead to a timeout in Vault’s auditing subroutine, potentially resulting in the Vault server to become unresponsive. This vulnerability, CVE-2025-6203, is fixed in Vault Community Edition 1.20.3 and Vault Enterprise 1.20.3, 1.19.9, 1.18.14, and 1.16.25.

A flaw was found in the GIF parser of GdkPixbuf’s LZW decoder. When an invalid symbol is encountered during decompression, the decoder sets the reported output size to the full buffer length rather than the actual number of written bytes. This logic error results in uninitialized sections of the buffer being included in the output, potentially leaking arbitrary memory contents in the processed image.

ssh in OpenSSH before 10.1 allows the '\0' character in an ssh:// URI, potentially leading to code execution when a ProxyCommand is used.

ssh in OpenSSH before 10.1 allows control characters in usernames that originate from certain possibly untrusted sources, potentially leading to code execution when a ProxyCommand is used. The untrusted sources are the command line and %-sequence expansion of a configuration file. (A configuration file that provides a complete literal username is not categorized as an untrusted source.)

An open redirect vulnerability has been identified in Grafana OSS organization switching functionality. Prerequisites for exploitation: - Multiple organizations must exist in the Grafana instance - Victim must be on a different organization than the one specified in the URL

Missing lock bit protection for NBIO registers could allow a local admin-privileged attacker to modify MMIO routing configurations, potentially resulting in loss of SEV-SNP guest integrity.

A flaw was found in libgepub, a library used to read EPUB files. The software mishandles file size calculations when opening specially crafted EPUB files, leading to incorrect memory allocations. This issue causes the application to crash. Known affected usage includes desktop services like Tumbler, which may process malicious files automatically when browsing directories. While no direct remote attack vectors are confirmed, any application using libgepub to parse user-supplied EPUB content could be vulnerable to a denial of service.

In fetchmail before 6.5.6, the SMTP client can crash when authenticating upon receiving a 334 status code in a malformed context.

A command injection vulnerability was discovered in the TrustyAI Explainability toolkit. Arbitrary commands placed in certain fields of a LMEValJob custom resource (CR) may be executed in the LMEvalJob pod's terminal. This issue can be exploited via a maliciously crafted LMEvalJob by a user with permissions to deploy a CR.

Happy DOM is a JavaScript implementation of a web browser without its graphical user interface. Happy DOM v19 and lower contains a security vulnerability that puts the owner system at the risk of RCE (Remote Code Execution) attacks. A Node.js VM Context is not an isolated environment, and if the user runs untrusted JavaScript code within the Happy DOM VM Context, it may escape the VM and get access to process level functionality. It seems like what the attacker can get control over depends on if the process is using ESM or CommonJS. With CommonJS the attacker can get hold of the `require()` function to import modules. Happy DOM has JavaScript evaluation enabled by default. This may not be obvious to the consumer of Happy DOM and can potentially put the user at risk if untrusted code is executed within the environment. Version 20.0.0 patches the issue by changing JavaScript evaluation to be disabled by default.

Sinatra is a domain-specific language for creating web applications in Ruby. In versions prior to 4.2.0, there is a denial of service vulnerability in the `If-Match` and `If-None-Match` header parsing component of Sinatra, if the `etag` method is used when constructing the response. Carefully crafted input can cause `If-Match` and `If-None-Match` header parsing in Sinatra to take an unexpected amount of time, possibly resulting in a denial of service attack vector. This header is typically involved in generating the `ETag` header value. Any applications that use the `etag` method when generating a response are impacted. Version 4.2.0 fixes the issue.

Authlib is a Python library which builds OAuth and OpenID Connect servers. Prior to version 1.6.5, Authlib’s JOSE implementation accepts unbounded JWS/JWT header and signature segments. A remote attacker can craft a token whose base64url‑encoded header or signature spans hundreds of megabytes. During verification, Authlib decodes and parses the full input before it is rejected, driving CPU and memory consumption to hostile levels and enabling denial of service. Version 1.6.5 patches the issue. Some temporary workarounds are available. Enforce input size limits before handing tokens to Authlib and/or use application-level throttling to reduce amplification risk.

Rack is a modular Ruby web server interface. Prior to versions 2.2.20, 3.1.18, and 3.2.3, `Rack::Request#POST` reads the entire request body into memory for `Content-Type: application/x-www-form-urlencoded`, calling `rack.input.read(nil)` without enforcing a length or cap. Large request bodies can therefore be buffered completely into process memory before parsing, leading to denial of service (DoS) through memory exhaustion. Users should upgrade to Rack version 2.2.20, 3.1.18, or 3.2.3, anu of which enforces form parameter limits using `query_parser.bytesize_limit`, preventing unbounded reads of `application/x-www-form-urlencoded` bodies. Additionally, enforce strict maximum body size at the proxy or web server layer (e.g., Nginx `client_max_body_size`, Apache `LimitRequestBody`).

OpenPrinting CUPS is an open source printing system for Linux and other Unix-like operating systems. Prior to version 2.4.15, a user in the lpadmin group can use the cups web ui to change the config and insert a malicious line. Then the cupsd process which runs as root will parse the new config and cause an out-of-bound write. This issue has been patched in version 2.4.15.

python-ldap is a lightweight directory access protocol (LDAP) client API for Python. In versions prior to 3.4.5, ldap.dn.escape_dn_chars() escapes \x00 incorrectly by emitting a backslash followed by a literal NUL byte instead of the RFC-4514 hex form \00. Any application that uses this helper to construct DNs from untrusted input can be made to consistently fail before a request is sent to the LDAP server (e.g., AD), resulting in a client-side denial of service. Version 3.4.5 contains a patch for the issue.

python-ldap is a lightweight directory access protocol (LDAP) client API for Python. In versions prior to 3.4.5, the sanitization method `ldap.filter.escape_filter_chars` can be tricked to skip escaping of special characters when a crafted `list` or `dict` is supplied as the `assertion_value` parameter, and the non-default `escape_mode=1` is configured. The method `ldap.filter.escape_filter_chars` supports 3 different escaping modes. `escape_mode=0` (default) and `escape_mode=2` happen to raise exceptions when a `list` or `dict` object is supplied as the `assertion_value` parameter. However, `escape_mode=1` computes without performing adequate logic to ensure a fully escaped return value. If an application relies on the vulnerable method in the `python-ldap` library to escape untrusted user input, an attacker might be able to abuse the vulnerability to launch ldap injection attacks which could potentially disclose or manipulate ldap data meant to be inaccessible to them. Version 3.4...

Improper Resource Shutdown or Release vulnerability in Apache Tomcat. If an error occurred (including exceeding limits) during the processing of a multipart upload, temporary copies of the uploaded parts written to disc were not cleaned up immediately but left for the garbage collection process to delete. Depending on JVM settings, application memory usage and application load, it was possible that space for the temporary copies of uploaded parts would be filled faster than GC cleared it, leading to a DoS. This issue affects Apache Tomcat: from 11.0.0-M1 through 11.0.11, from 10.1.0-M1 through 10.1.46, from 9.0.0.M1 through 9.0.109. 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 11.0.12 or later, 10.1.47 or later or 9.0.110 or later which fixes the issue.