In Mesa before 25.3.6 and 26 before 26.0.1, out-of-bounds memory access can occur in WebGPU because the amount of to-be-allocated data depends on an untrusted party, and is then used for alloca.

LiteLLM is a proxy server (AI Gateway) to call LLM APIs in OpenAI (or native) format. From version 1.81.16 to before version 1.83.7, a database query used during proxy API key checks mixed the caller-supplied key value into the query text instead of passing it as a separate parameter. An unauthenticated attacker could send a specially crafted Authorization header to any LLM API route (for example POST /chat/completions) and reach this query through the proxy's error-handling path. An attacker could read data from the proxy's database and may be able to modify it, leading to unauthorised access to the proxy and the credentials it manages. This issue has been patched in version 1.83.7.

Axios is a promise based HTTP client for the browser and Node.js. From 1.0.0 to before 1.15.1, the FormDataPart constructor in lib/helpers/formDataToStream.js interpolates value.type directly into the Content-Type header of each multipart part without sanitizing CRLF (\r\n) sequences. An attacker who controls the .type property of a Blob/File-like object (e.g., via a user-uploaded file in a Node.js proxy service) can inject arbitrary MIME part headers into the multipart form-data body. This bypasses Node.js v18+ built-in header protections because the injection targets the multipart body structure, not HTTP request headers. This vulnerability is fixed in 1.15.1.

A code injection in Ivanti Endpoint Manager Mobile allowing attackers to achieve unauthenticated remote code execution.

cPanel and WHM versions after 11.40 contain an authentication bypass vulnerability in the login flow that allows unauthenticated remote attackers to gain unauthorized access to the control panel.

Improper neutralization of inputs used in an OS command in the FSx Windows File Server volume mounting component in Amazon ECS Agent on Windows before version 1.103.0 might allow a remote authenticated threat actor to execute shell commands with SYSTEM privileges on the underlying host via a specially crafted username field in an ECS task definition. This issue requires permissions to register ECS task definitions or write to the Secrets Manager or SSM Parameter Store credentials used by the FSx volume configuration. To remediate this issue, users should upgrade to version 1.103.0.

Pathological inputs could cause DoS through consumePhrase when parsing an email address according to RFC 5322.

If a trusted template author were to write a <script> tag containing an empty 'type' attribute or a 'type' attribute with an ASCII whitespace, the execution of the template would incorrectly escape any data passed into the <script> block.

Akamai Guardicore Platform Agent (GPA) and Zero Trust Client on Linux and macOS allow TOCTOU-based local privilege escalation. The GPA service creates an IPC socket in the world-writable /tmp directory. It accepts unauthenticated IPC control messages. This enables a TOCTOU vulnerability in the HandleSaveLogs() function of the GPA service, by creating a log file and manipulating it into a symlink that points to the targeted path; this can allow an unprivileged local user to make arbitrary root-owned files world-writable. In addition, a diagnostic collection tool (gimmelogs) running with root privileges was vulnerable to command injection from the dbstore, offering a second privilege escalation vector. (On Windows, gimmelogs does not have command injection but does allow writing a ZIP archive to an unintended location.) This affects Akamai Guardicore Platform Agent 7.0 through 7.3.1 and Akamai Zero Trust Client 6.0 through 6.1.5.

Well-crafted inputs reaching ParseAddress, ParseAddressList, and ParseDate were able to trigger excessive CPU exhaustion and memory allocations.