vLLM is an inference and serving engine for large language models (LLMs). From 0.3.0 until 0.22.0, a vulnerability in ASGI web servers and starlette's trust on those web servers enables an authentication bypass of the OpenAI API AuthenticationMiddleware. It allows to use the API without providing the configured VLLM_API_KEY or --api-key. This vulnerability is fixed in 0.22.0.

Angular is a development platform for building mobile and desktop web applications using TypeScript/JavaScript and other languages. Prior to 22.0.0-rc.2, 21.2.15, 20.3.22, and 19.2.23, an issue in the @angular/service-worker package compromises the integrity of request-policy enforcement during request reconstruction. When the Angular Service Worker intercepts network requests for matched assets, it reconstructs a new Request object using an internal helper function. During this reconstruction process, the helper function strips explicit client-defined safety parameters: the credentials configuration (such as credentials: 'omit') and the HTTP cache mode configuration (such as cache: 'no-store'). These are reverted back to standard browser-default parameters (credentials: 'same-origin' and default HTTP cache properties). This causes the browser to include active credentials (such as cookies or Authorization headers) on outbound requests where the client-side developer explicitly instructed they should be omitted, leading to potential session leaks. Additionally, it causes private or non-cacheable resources to be cached by the service worker's engine, making private page states accessible or persistent inside the client's local cache post-logout. This vulnerability is fixed in 22.0.0-rc.2, 21.2.15, 20.3.22, and 19.2.23.

vLLM is an inference and serving engine for large language models (LLMs). Prior to 0.22.0, vLLM's revision pinning controls do not consistently apply to all artifacts loaded for a model. A deployment that supplies --revision or --code-revision can still load dynamic code, GGUF files, image processors, retrieval side weights, or same-repository subfolder weights/config from an unpinned/default revision. This is a supply-chain integrity issue for pinned vLLM deployments. Operators can believe they are serving a reviewed model revision while vLLM resolves behavior-affecting nested or sibling artifacts outside that reviewed revision. This vulnerability is fixed in 0.22.0.

Angular is a development platform for building mobile and desktop web applications using TypeScript/JavaScript and other languages. Prior to 22.0.0-rc.2, 21.2.15, 20.3.22, and 19.2.23, a Denial of Service (DoS) vulnerability exists in the @angular/common package of Angular. The formatNumber function, which is also utilized by DecimalPipe, PercentPipe, and CurrencyPipe, does not properly validate the upper bounds of the digitsInfo parameter. Specifically, the minimum and maximum fraction digits parsed from the digitsInfo string (e.g., 1.2-4) are converted to integers and used without limits. When parsing a maliciously crafted digitsInfo string with excessively large fraction digit values (e.g., 1.200000000-200000000), the internal roundNumber function attempts to pad the digits array to match the requested fraction size. This results in an unbounded loop that repeatedly pushes elements into an array. This vulnerability is fixed in 22.0.0-rc.2, 21.2.15, 20.3.22, and 19.2.23.

The Angular Language Service VS Code Extension provides a rich editing experience for Angular templates. the client-side Angular Language Service VS Code extension configures the tooltip Markdown renderer with the isTrusted: true option (located in client/src/client.ts). This setting instructs VS Code to trust all rendered content it receives, which enables active elements such as command: URIs. However, the background Angular Language Server process fails to escape or sanitize brackets, raw links, and control characters from JSDoc strings before forwarding the hover Markdown content (located in server/src/handlers/hover.ts and server/src/text_render.ts). An attacker can leverage this behavior by crafting a project TypeScript or JavaScript file (or a third-party npm package dependency) containing a malicious JSDoc tooltip with an embedded active command link. When a developer hovers over the target symbol to render the tooltip and clicks the malicious link, the IDE executes the command sequence directly on the developer's host machine. Prior to 21.2.4, This vulnerability is fixed in 21.2.4.

Angular is a development platform for building mobile and desktop web applications using TypeScript/JavaScript and other languages. Prior to 22.0.1, 21.2.17, and 20.3.25, to optimize client-side bootstrap in Server-Side Rendered (SSR) environments, Angular supports Hydration via provideClientHydration(). During SSR, Angular serializes the application's runtime state (such as cached HttpClient responses) and outputs it into the HTML stream as a <script> tag with a predictable identifier. During client bootstrap, Angular recovers this state by looking up the element via document.getElementById('ng-state') and parsing its text content. Because the DOM element lookup for the state container is predictable and relies solely on the ID selector (ng-state), it is susceptible to DOM Clobbering. If the application binds untrusted user input or CMS content to element properties such as id (e.g., <div [id]="userInput"> or <a id="ng-state">) before the genuine <script> tag is parsed by the browser, the attacker-controlled element takes precedence in the DOM lookup. During hydration, when Angular calls document.getElementById('ng-state'), the browser returns the attacker's clobbered element. Angular then attempts to parse the text content or attributes of this clobbered element as JSON. This vulnerability is fixed in 22.0.1, 21.2.17, and 20.3.25.

Python-Multipart is a streaming multipart parser for Python. Prior to 0.0.31, parse_form() did not validate the Content-Length header before using it to bound its chunked read of the request body. A negative Content-Length turned the bounded read into a read-until-EOF, so the entire body was loaded into memory in a single read instead of in fixed-size chunks. This vulnerability is fixed in 0.0.31.

js-yaml is a JavaScript YAML parser and dumper. Prior to 4.2.0, a crafted YAML document can trigger algorithmic CPU exhaustion in js-yaml merge-key processing (<<) by repeating the same alias many times in a merge sequence. This causes quadratic parse-time behavior relative to input size and can block a Node.js worker/event loop for seconds with a relatively small payload (tens of KB), resulting in denial of service. The issue is in merge handling inside lib/loader.js. This vulnerability is fixed in 4.2.0.

NLTK (Natural Language Toolkit) is a suite of open source Python modules, data sets, and tutorials supporting research and development in Natural Language Processing. Prior to 3.10.0-rc1, nltk.data.load() in NLTK is vulnerable to path traversal via URL-encoded path separators and traversal segments when using the nltk: URL scheme. The unsafe-path regex check is performed before url2pathname() decodes the %xx sequences (a classic decode-after-check / TOCTOU-style flaw), allowing an attacker to bypass the protection documented in NLTK's SECURITY.md and read arbitrary files from the filesystem. While literal traversal strings such as ../../../etc/passwd are correctly blocked, encoded variants such as %2fetc%2fpasswd, %2e%2e%2f..., and ..%2f..%2f slip past the regex and are subsequently decoded into a real filesystem path. This vulnerability is fixed in 3.10.0-rc1.

A vulnerability was determined in BerriAI litellm up to 1.82.2. This affects the function json.dumps of the file litellm/proxy/management_endpoints/ui_sso.py of the component SSO Debug Flow. Executing a manipulation can lead to missing authentication. The attack can be executed remotely. The exploit has been publicly disclosed and may be utilized. The vendor was contacted early about this disclosure.