Exploit Intelligence Platform

ReDoS in strip_whitespaces() function in cps/string_helper.py in Calibre Web and Autocaliweb allows unauthenticated remote attackers to cause denial of service via specially crafted username parameter that triggers catastrophic backtracking during login. This issue affects Calibre Web: 0.6.24 (Nicolette); Autocaliweb: from 0.7.0 before 0.7.1.

LLama-Index CLI version v0.12.20 contains an OS command injection vulnerability. The vulnerability arises from the improper handling of the `--files` argument, which is directly passed into `os.system`. An attacker who controls the content of this argument can inject and execute arbitrary shell commands. This vulnerability can be exploited locally if the attacker has control over the CLI arguments, and remotely if a web application calls the LLama-Index CLI with a user-controlled filename. This issue can lead to arbitrary code execution on the affected system.

vLLM, an inference and serving engine for large language models (LLMs), has an issue in versions 0.6.5 through 0.8.4 that ONLY impacts environments using the `PyNcclPipe` KV cache transfer integration with the V0 engine. No other configurations are affected. vLLM supports the use of the `PyNcclPipe` class to establish a peer-to-peer communication domain for data transmission between distributed nodes. The GPU-side KV-Cache transmission is implemented through the `PyNcclCommunicator` class, while CPU-side control message passing is handled via the `send_obj` and `recv_obj` methods on the CPU side.​ The intention was that this interface should only be exposed to a private network using the IP address specified by the `--kv-ip` CLI parameter. The vLLM documentation covers how this must be limited to a secured network. The default and intentional behavior from PyTorch is that the `TCPStore` interface listens on ALL interfaces, regardless of what IP address is provided. The IP address given was only used as a client-side address to use. vLLM was fixed to use a workaround to force the `TCPStore` instance to bind its socket to a specified private interface. As of version 0.8.5, vLLM limits the `TCPStore` socket to the private interface as configured.

h11 is a Python implementation of HTTP/1.1. Prior to version 0.16.0, a leniency in h11's parsing of line terminators in chunked-coding message bodies can lead to request smuggling vulnerabilities under certain conditions. This issue has been patched in version 0.16.0. Since exploitation requires the combination of buggy h11 with a buggy (reverse) proxy, fixing either component is sufficient to mitigate this issue.

vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. When vLLM is configured to use Mooncake, unsafe deserialization exposed directly over ZMQ/TCP on all network interfaces will allow attackers to execute remote code on distributed hosts. This is a remote code execution vulnerability impacting any deployments using Mooncake to distribute KV across distributed hosts. This vulnerability is fixed in 0.8.0.

An issue was discovered in Django 4.2 before 4.2.21, 5.1 before 5.1.9, and 5.2 before 5.2.1. The django.utils.html.strip_tags() function is vulnerable to a potential denial-of-service (slow performance) when processing inputs containing large sequences of incomplete HTML tags. The template filter striptags is also vulnerable, because it is built on top of strip_tags().

LNbits is a Lightning wallet and accounts system. A Server-Side Request Forgery (SSRF) vulnerability has been discovered in LNbits' LNURL authentication handling functionality. When processing LNURL authentication requests, the application accepts a callback URL parameter and makes an HTTP request to that URL using the httpx library with redirect following enabled. The application doesn't properly validate the callback URL, allowing attackers to specify internal network addresses and access internal resources.

When running the AWS Serverless Application Model Command Line Interface (SAM CLI) build process with Docker and symlinks are included in the build files, the container environment allows a user to access privileged files on the host by leveraging the elevated permissions granted to the tool. A user could leverage the elevated permissions to access restricted files via symlinks and copy them to a more permissive location on the container. Users should upgrade to v1.133.0 or newer and ensure any forked or derivative code is patched to incorporate the new fixes.

Label Studio is an open source data labeling tool. Prior to version 1.16.0, Label Studio's `/projects/upload-example` endpoint allows injection of arbitrary HTML through a `GET` request with an appropriately crafted `label_config` query parameter. By crafting a specially formatted XML label config with inline task data containing malicious HTML/JavaScript, an attacker can achieve Cross-Site Scripting (XSS). While the application has a Content Security Policy (CSP), it is only set in report-only mode, making it ineffective at preventing script execution. The vulnerability exists because the upload-example endpoint renders user-provided HTML content without proper sanitization on a GET request. This allows attackers to inject and execute arbitrary JavaScript in victims' browsers by getting them to visit a maliciously crafted URL. This is considered vulnerable because it enables attackers to execute JavaScript in victims' contexts, potentially allowing theft of sensitive data, session hijacking, or other malicious actions. Version 1.16.0 contains a patch for the issue.

man-group/dtale version 3.10.0 is vulnerable to an authentication bypass and remote code execution (RCE) due to improper input validation. The vulnerability arises from a hardcoded `SECRET_KEY` in the flask configuration, allowing attackers to forge a session cookie if authentication is enabled. Additionally, the application fails to properly restrict custom filter queries, enabling attackers to execute arbitrary code on the server by bypassing the restriction on the `/update-settings` endpoint, even when `enable_custom_filters` is not enabled. This vulnerability allows attackers to bypass authentication mechanisms and execute remote code on the server.

A remote code execution vulnerability exists in invoke-ai/invokeai versions 5.3.1 through 5.4.2 via the /api/v2/models/install API. The vulnerability arises from unsafe deserialization of model files using torch.load without proper validation. Attackers can exploit this by embedding malicious code in model files, which is executed upon loading. This issue is fixed in version 5.4.3.

PyMySQL through 1.1.0 allows SQL injection if used with untrusted JSON input because keys are not escaped by escape_dict.

Jinja is an extensible templating engine. The `xmlattr` filter in affected versions of Jinja accepts keys containing non-attribute characters. XML/HTML attributes cannot contain spaces, `/`, `>`, or `=`, as each would then be interpreted as starting a separate attribute. If an application accepts keys (as opposed to only values) as user input, and renders these in pages that other users see as well, an attacker could use this to inject other attributes and perform XSS. The fix for CVE-2024-22195 only addressed spaces but not other characters. Accepting keys as user input is now explicitly considered an unintended use case of the `xmlattr` filter, and code that does so without otherwise validating the input should be flagged as insecure, regardless of Jinja version. Accepting _values_ as user input continues to be safe. This vulnerability is fixed in 3.1.4.

A directory traversal vulnerability exists in the zenml-io/zenml repository, specifically within the /api/v1/steps endpoint. Attackers can exploit this vulnerability by manipulating the 'logs' URI path in the request to fetch arbitrary file content, bypassing intended access restrictions. The vulnerability arises due to the lack of validation for directory traversal patterns, allowing attackers to access files outside of the restricted directory.

In eosphoros-ai/db-gpt version v0.6.0, the web API `POST /api/v1/editor/sql/run` allows execution of arbitrary SQL queries without any access control. This vulnerability can be exploited by attackers to perform Arbitrary File Write using DuckDB SQL, enabling them to write arbitrary files to the victim's file system. This can potentially lead to Remote Code Execution (RCE).

aiohttp is an asynchronous HTTP client/server framework for asyncio and Python. When using aiohttp as a web server and configuring static routes, it is necessary to specify the root path for static files. Additionally, the option 'follow_symlinks' can be used to determine whether to follow symbolic links outside the static root directory. When 'follow_symlinks' is set to True, there is no validation to check if reading a file is within the root directory. This can lead to directory traversal vulnerabilities, resulting in unauthorized access to arbitrary files on the system, even when symlinks are not present. Disabling follow_symlinks and using a reverse proxy are encouraged mitigations. Version 3.9.2 fixes this issue.

A remote code execution (RCE) vulnerability exists in the lightning-ai/pytorch-lightning library version 2.2.1 due to improper handling of deserialized user input and mismanagement of dunder attributes by the `deepdiff` library. The library uses `deepdiff.Delta` objects to modify application state based on frontend actions. However, it is possible to bypass the intended restrictions on modifying dunder attributes, allowing an attacker to construct a serialized delta that passes the deserializer whitelist and contains dunder attributes. When processed, this can be exploited to access other modules, classes, and instances, leading to arbitrary attribute write and total RCE on any self-hosted pytorch-lightning application in its default configuration, as the delta endpoint is enabled by default.

changedetection.io is an open source web page change detection, website watcher, restock monitor and notification service. There is a Server Side Template Injection (SSTI) in Jinja2 that allows Remote Command Execution on the server host. Attackers can run any system command without any restriction and they could use a reverse shell. The impact is critical as the attacker can completely takeover the server machine. This can be reduced if changedetection is behind a login page, but this isn't required by the application (not by default and not enforced).

Werkzeug is a comprehensive WSGI web application library. The debugger in affected versions of Werkzeug can allow an attacker to execute code on a developer's machine under some circumstances. This requires the attacker to get the developer to interact with a domain and subdomain they control, and enter the debugger PIN, but if they are successful it allows access to the debugger even if it is only running on localhost. This also requires the attacker to guess a URL in the developer's application that will trigger the debugger. This vulnerability is fixed in 3.0.3.

pgAdmin <= 8.4 is affected by a Remote Code Execution (RCE) vulnerability through the validate binary path API. This vulnerability allows attackers to execute arbitrary code on the server hosting PGAdmin, posing a severe risk to the database management system's integrity and the security of the underlying data.