Schwachstellen im Stack Python · Seite 151

Arbitrary Code execution in Keras Safe Mode

The Keras Model.load_model method can be exploited to achieve arbitrary code execution, even with safe_mode=True. One can create a specially crafted .keras model archive that, when loaded via Model.load_model, will trigger arbitrary code to be executed. This is achieved by crafting a special config.json (a file within the .keras archive) that will invoke keras.config.enable_unsafe_deserialization() to disable safe mode. Once safe mode is disable, one can use the Lambda layer feature of keras, which allows arbitrary Python code in the form of pickled code. Both can appear in the same archive. Simply the keras.config.enable_unsafe_deserializat

• Auf 3.11.0 aktualisieren

Path Traversal and Arbitrary File Deletion in invoke-ai/invokeai

A vulnerability in invokeai version v6.0.0a1 and below allows attackers to perform path traversal and arbitrary file deletion via the GET /api/v1/images/download/{bulk_download_item_name} endpoint. By manipulating the filename arguments, attackers can read and delete any files on the server, including critical system files such as SSH keys, databases, and configuration files. This vulnerability results in high confidentiality, integrity, and availability impacts.

• Auf 6.7.0 aktualisieren

PickleScan Bypasses Unsafe Globals Check Using Submodule Imports

A Protection Mechanism Failure vulnerability in mmaitre314 picklescan versions up to and including 0.0.30 allows a remote attacker to bypass the unsafe globals check. This is possible because the scanner performs an exact match for module names, allowing malicious payloads to be loaded via submodules of dangerous packages (e.g., 'asyncio.unix_events' instead of 'asyncio'). When the incorrectly considered safe file is loaded after scan, it can lead to the execution of malicious code.

• Auf 0.0.31 aktualisieren

PickleScan Security Bypass via Bad CRC in ZIP Archive

An Improper Handling of Exceptional Conditions vulnerability in the ZIP archive scanning component of mmaitre314 picklescan allows a remote attacker to bypass security scans. This is achieved by crafting a ZIP archive containing a file with a bad Cyclic Redundancy Check (CRC), which causes the scanner to halt and fail to analyze the contents for malicious pickle files. When the file incorrectly considered safe is loaded, it can lead to the execution of malicious code.

• Auf 0.0.31 aktualisieren

PickleScan Security Bypass Using Misleading File Extension

An Improper Input Validation vulnerability in the scanning logic of mmaitre314 picklescan versions up to and including 0.0.30 allows a remote attacker to bypass pickle files security checks by supplying a standard pickle file with a PyTorch-related file extension. When the pickle file incorrectly considered safe is loaded, it can lead to the execution of malicious code.

• Auf 0.0.31 aktualisieren

CVE-2025-59377

feiskyer mcp-kubernetes-server through 0.1.11 allows OS command injection, even in read-only mode, via /mcp/kubectl because shell=True is used. NOTE: this is unrelated to mcp-server-kubernetes and CVE-2025-53355.

CVE-2025-59376

feiskyer mcp-kubernetes-server through 0.1.11 does not consider chained commands in the implementation of --disable-write and --disable-delete, e.g., it allows a "kubectl version; kubectl delete pod" command because the first word (i.e., "version") is not a write or delete operation.

CVE-2025-49089

wangxutech MoneyPrinterTurbo 1.2.6 allows path traversal via /api/v1/download/ URIs such as /api/v1/download//etc/passwd.

Regular Expression Denial of Service (ReDoS) in huggingface/transformers

A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically within the `normalize_numbers()` method of the `EnglishNormalizer` class. This vulnerability affects versions up to 4.52.4 and is fixed in version 4.53.0. The issue arises from the method's handling of numeric strings, which can be exploited using crafted input strings containing long sequences of digits, leading to excessive CPU consumption. This vulnerability impacts text-to-speech and number normalization tasks, potentially causing service disruption, resource exhaustion, and API vulnerabilities.

• Auf 4.53.0 aktualisieren

Regular Expression Denial of Service (ReDoS) in huggingface/transformers

A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically affecting the MarianTokenizer's `remove_language_code()` method. This vulnerability is present in version 4.52.4 and has been fixed in version 4.53.0. The issue arises from inefficient regex processing, which can be exploited by crafted input strings containing malformed language code patterns, leading to excessive CPU consumption and potential denial of service.

• Auf 4.53.0 aktualisieren

Flask App Builder has an Authentication Bypass vulnerability when using non AUTH_DB methods

Flask-AppBuilder is an application development framework. Prior to version 4.8.1, when Flask-AppBuilder is configured to use OAuth, LDAP, or other non-database authentication methods, the password reset endpoint remains registered and accessible, despite not being displayed in the user interface. This allows an enabled user to reset their password and be able to create JWT tokens even after the user is disabled on the authentication provider. Users should upgrade to Flask-AppBuilder version 4.8.1 or later to receive a fix. If immediate upgrade is not possible, manually disable password reset routes in the application configuration; implement

• Auf 4.8.1 aktualisieren

Neo4j Cypher MCP server is vulnerable to DNS rebinding attacks

DNS rebinding vulnerability in Neo4j Cypher MCP server allows malicious websites to bypass Same-Origin Policy protections and execute unauthorised tool invocations against locally running Neo4j MCP instances. The attack relies on the user being enticed to visit a malicious website and spend sufficient time there for DNS rebinding to succeed.

• Auf 0.4.0 aktualisieren

Indico vulnerable to Cross-Site Scripting via LaTeX math code

Indico is an event management system that uses Flask-Multipass, a multi-backend authentication system for Flask. Prior to version 3.3.8, there is a Cross-Site-Scripting vulnerability when rendering LaTeX math code in contribution or abstract descriptions. Users should to update to Indico 3.3.8 as soon as possible. As a workaround, only let trustworthy users create content on Indico. Note that a conference doing a Call for Abstracts actively invites external speakers (who the organizers may not know and thus cannot fully trust) to submit content, hence the need to update to a a fixed version ASAP in particular when using such workflows.

• Auf 3.3.8 aktualisieren

Indico may disclose unauthorized user details access via legacy API

Indico is an event management system that uses Flask-Multipass, a multi-backend authentication system for Flask. Prior to version 3.3.8, a legacy API to retrieve user details could be misused to retrieve profile details of other users without having admin permissions due to a broken access check. Users should to update to Indico 3.3.8 as soon as possible. As a workaround, it is possible to restrict access to the affected API (e.g. in the webserver config).

• Auf 3.3.8 aktualisieren

PyInstaller has local privilege escalation vulnerability

PyInstaller bundles a Python application and all its dependencies into a single package. Due to a special entry being appended to `sys.path` during the bootstrap process of a PyInstaller-frozen application, and due to the bootstrap script attempting to load an optional module for bytecode decryption while this entry is still present in `sys.path`, an application built with PyInstaller < 6.0.0 may be tricked by an unprivileged attacker into executing arbitrary python code when **all** of the following conditions are met. First, the application is built with PyInstaller < 6.0.0; both onedir and onefile mode are affected. Second, the optional by

• Auf 6.0.0 aktualisieren

Infrahub allows authentication with deleted and expired API tokens

Infrahub offers a central hub to manage data, templates, and playbooks. Prior to versiond 1.3.9 and 1.4.5, a bug in the authentication logic will cause API tokens that were deleted and/or expired to be considered valid. This means that any API token that is associated with an active user account can authenticate successfully. This issue is fixed in versions 1.3.9 and 1.4.5. As a workaround, users can delete or deactivate the account associated with a deleted API token to prevent that token from authenticating.

• Auf 1.3.9 aktualisieren
• Auf 1.4.5 aktualisieren

copyparty: Sharing a single file does not fully restrict access to other files in source folder

Copyparty is a portable file server. In versions prior to 1.19.8, there was a missing permission-check in the shares feature (the `shr` global-option). When a share was created for just one file inside a folder, it was possible to access the other files inside that folder by guessing the filenames. It was not possible to descend into subdirectories in this manner; only the sibling files were accessible. This issue did not affect filekeys or dirkeys. Version 1.19.8 fixes the issue.

• Auf 1.19.8 aktualisieren

OctoPrint is Vulnerable to RCE Attacks via Unsanitized Filename in File Upload

OctoPrint provides a web interface for controlling consumer 3D printers. OctoPrint versions up until and including 1.11.2 contain a vulnerability that allows an authenticated attacker to upload a file under a specially crafted filename that will allow arbitrary command execution if said filename becomes included in a command defined in a system event handler and said event gets triggered. If no event handlers executing system commands with uploaded filenames as parameters have been configured, this vulnerability does not have an impact. The vulnerability is patched in version 1.11.3. As a workaround, OctoPrint administrators who have event ha

• Auf 1.11.3 aktualisieren

lmsys sglang update_weights_from_tensor main deserialization

A security flaw has been discovered in lmsys sglang 0.4.6. Affected by this vulnerability is the function main of the file /update_weights_from_tensor. The manipulation of the argument serialized_named_tensors results in deserialization. The attack can be launched remotely. The exploit has been released to the public and may be exploited. The vendor was contacted early about this disclosure but did not respond in any way.

• Auf 0.5.4 aktualisieren

MONAI's unsafe use of Pickle deserialization may lead to RCE

MONAI (Medical Open Network for AI) is an AI toolkit for health care imaging. In versions up to and including 1.5.0, the `pickle_operations` function in `monai/data/utils.py` automatically handles dictionary key-value pairs ending with a specific suffix and deserializes them using `pickle.loads()` . This function also lacks any security measures. The deserialization may lead to code execution. As of time of publication, no known fixed versions are available.

• Auf 1.5.1 aktualisieren
• Auf 1.5.1rc1 aktualisieren