Schwachstellen im Stack Python · Seite 89

Arbitrary Remote Code Execution via `_attn_implementation_internal` Config Injection in huggingface/transformers

A critical remote code execution vulnerability exists in all versions of the HuggingFace transformers library prior to version 5.3.0. The vulnerability allows an attacker to craft a malicious `config.json` file containing the `_attn_implementation_internal` field set to an attacker-controlled HuggingFace Hub repository ID. When a victim loads this model using the standard `AutoModelForCausalLM.from_pretrained()` API, the library downloads and executes arbitrary Python code from the attacker's repository with the victim's full OS privileges. This issue arises due to unfiltered deserialization of configuration attributes, insufficient sanitizat

• Auf 5.3.0 aktualisieren

NousResearch hermes-agent CLI web-dashboard web_server.py _discover_dashboard_plugins comparison

A security flaw has been discovered in NousResearch hermes-agent 2026.4.23. Affected is the function _discover_dashboard_plugins of the file hermes_cli/web_server.py of the component CLI web-dashboard Interface. Performing a manipulation of the argument HERMES_ENABLE_PROJECT_PLUGINS results in incorrect comparison. The attack is only possible with local access. The exploit has been released to the public and may be used for attacks. The vendor was contacted early about this disclosure but did not respond in any way.

• Auf 0.15.0 aktualisieren

NousResearch hermes-agent Environment Variable code_execution_tool.py execute_code sandbox

A vulnerability was identified in NousResearch hermes-agent up to 2026.4.16. This impacts the function execute_code of the file tools/code_execution_tool.py of the component Environment Variable Handler. Such manipulation leads to sandbox issue. It is possible to launch the attack remotely. The exploit is publicly available and might be used. The vendor was contacted early about this disclosure but did not respond in any way.

• Auf 0.11.0 aktualisieren

NousResearch hermes-agent prompt_builder.py _scan_context_content injection

A vulnerability was found in NousResearch hermes-agent 2026.4.23. The impacted element is the function _scan_context_content of the file agent/prompt_builder.py. The manipulation results in injection. The attack may be performed from remote. The exploit has been made public and could be used. The vendor was contacted early about this disclosure but did not respond in any way.

• Auf 0.15.0 aktualisieren

NousResearch hermes-agent Skills Guard Multi-Word Prompt skills_guard.py injection

A security vulnerability has been detected in NousResearch hermes-agent up to 2026.4.23. Impacted is an unknown function of the file agent/skills_guard.py of the component Skills Guard Multi-Word Prompt Handler. The manipulation of the argument THREAT_PATTERNS leads to injection. Remote exploitation of the attack is possible. The exploit has been disclosed publicly and may be used. The vendor was contacted early about this disclosure but did not respond in any way.

• Auf 0.15.0 aktualisieren

Argument Injection in prefecthq/prefect

A vulnerability in the `GitHubRepository` block of the `prefect-github` integration in Prefect version 3.6.18 allows an attacker to inject arbitrary git command-line options via the `reference` field. The `reference` field is concatenated directly into a `git clone` command string without proper sanitization, and then parsed by `shlex.split()`. This enables injection of options such as `-c`, leading to potential Server-Side Request Forgery (SSRF), credential theft, or remote code execution (RCE). The vulnerability affects both the `aget_directory()` and `get_directory()` methods in `src/integrations/prefect-github/prefect_github/repository.py

instagrapi: Unsafe signup challenge path handling in instagrapi

instagrapi versions before 2.6.9 accepted server-supplied signup challenge paths and used them to build request URLs before validating that the paths were relative Instagram API paths. A malicious or tampered challenge payload could cause challenge handling requests to be sent outside the intended Instagram host with the client\'s existing session headers. Version 2.6.9 validates challenge paths before building URLs, solving captcha challenges, or submitting phone/SMS challenge forms.

• Auf 2.6.9 aktualisieren

JupyterHub: Cross-origin form POSTs bypass XSRF

JupyterHub is software that allows users to create a multi-user server for Jupyter notebooks. In versions 4.1.0 through 5.4.4, XSRF protection (updated in 4.1.0) inappropriately treated requests with Sec-Fetch-Mode: no-cors as same-origin requests, bypassing XSRF checks. The JSON API is not affected, only HTTP form endpoints, such as /hub/spawn and /hub/accept-share, meaning attackers could trigger server spawn (but not access the server) and if the attacker is a JupyterHub user permitted to share access to their server, cause a user to accept a share and have access to the attacker's server. This issue has been fixed in version 5.4.5. If dev

• Auf 5.4.5 aktualisieren

BentoML has Information Disclosure in `bentoml build` via symlink traversal in the build context

BentoML is a Python library for building online serving systems optimized for AI apps and model inference. In versions 1.4.38 and prior, the build packaging workflow follows attacker-controlled symlinks inside the build context and copies the referenced file contents into the generated Bento artifact. If a victim builds an untrusted repository or other attacker-supplied build context, the attacker can place a symlink such as loot.txt -> /tmp/outside-marker.txt or a link to a more sensitive local file. When bentoml build runs, BentoML dereferences the symlink and packages the target file contents into the Bento. The leaked file can then propag

• Auf 1.4.39 aktualisieren

Insecure Deserialization in Amazon Braket SDK Job Results Processing

Insecure deserialization in the job results processing component in Amazon Braket SDK before 1.117.0 might allow a remote authenticated user with S3 write access to the job output bucket to achieve arbitrary code execution on any machine that processes job results. We recommend you upgrade to amazon-braket-sdk version 1.117.0 or later.

• Auf 1.117.0 aktualisieren

LiteLLM < 1.83.10 Privilege Escalation via User Update

LiteLLM prior to 1.83.10 allows a user to modify their own user_role via the /user/update endpoint. While the endpoint correctly restricts users to updating only their own account, it does not restrict which fields may be changed. A user who can reach this endpoint can set their role to proxy_admin, gaining full administrative access to LiteLLM including all users, teams, keys, models, and prompt history. Users with the org_admin role have legitimate access to this endpoint and can exploit this vulnerability without chaining any additional flaw.

• Auf 1.83.10 aktualisieren

LiteLLM < 1.83.14 Privilege Escalation via API Key Generation

LiteLLM prior to 1.83.14 allows an authenticated internal_user to create API keys with access to routes that their role does not permit. When generating a key, the allowed_routes field is stored without verifying that the specified routes fall within the user's own permissions. A key created with access to admin-only routes can then be used to reach those routes successfully, bypassing the role-based access controls that would otherwise block the request, enabling full privilege escalation from internal_user to proxy_admin.

• Auf 1.83.14 aktualisieren

Apache Fory: PyFory ReduceSerializer Incomplete Policy Enforcement

Deserialization of untrusted data in Apache Fory PyFory. PyFory's ReduceSerializer could bypass documented DeserializationPolicy validation hooks during reduce-state restoration and global-name resolution. An application is vulnerable if it deserializes attacker-controlled data using PyFory Python-native mode with strict mode disabled and relies on DeserializationPolicy to restrict unsafe classes, functions, or module attributes. This issue affects Apache Fory: from before 1.0.0. Mitigation: Users of Apache Fory are recommended to upgrade to version 1.0.0 or later, which enforces DeserializationPolicy validation for the affected ReduceSeria

• Auf 1.0.0 aktualisieren

Authorization Bypass in SearchModelVersions in mlflow/mlflow

In mlflow/mlflow versions up to 3.9.0, the `SearchModelVersions` REST API endpoint and the `mlflowSearchModelVersions` GraphQL query lack proper per-model authorization checks when basic authentication is enabled. This allows any authenticated user to enumerate all model versions across all registered models, regardless of their permission level. The issue arises due to the absence of `SearchModelVersions` in the `BEFORE_REQUEST_VALIDATORS` and `AFTER_REQUEST_HANDLERS` for the REST API, and its omission from `GraphQLAuthorizationMiddleware.PROTECTED_FIELDS` for GraphQL. This vulnerability can expose sensitive information such as model names,

• Auf 3.10.0 aktualisieren

wger: cross-tenant account deletion / deactivation / activation by gym.manage_gym + gym=None

## Summary GHSA-mhc8-p3jx-84mm (CVE-2026-43948) reported that wger's `reset_user_password` and `gym_permissions_user_edit` views in `wger/gym/views/user.py` performed a gym-scope authorization check using Django ORM object comparison (`if request.user.userprofile.gym != user.userprofile.gym`) which silently passes when both sides are `None` (`None != None` evaluates to `False`). The maintainer's suggested patch ("Apply the same `same_gym()` helper pattern to all five views sharing this check") replaces every `userprofile.gym !=` site with the new `is_same_gym()` helper that explicitly excludes `None` (`gym_a is not None and gym_a == gym_b`).

Apache Airflow CNCF Kubernetes provider: JWT Token Exposure in KubernetesExecutor Command-Line Arguments

JWT tokens that were used by workers in Kubernetes Executors have been exposed to users who had read only access to Kuberentes Pods. This could allow users with just read-only access to perform actions that were only available to running tasks via Task SDK and potentially allow to modify state of Airflow Database for tasks.

• Auf 10.17.0 aktualisieren

Apache Airflow Amazon provider: Prevent unauthorized access to team-scoped secrets in AWS Secrets Manager and SSM Parameter Store backends

In the AWS Secrets Manager and SSM Parameter Store secrets backends of `apache-airflow-providers-amazon` prior to 9.28.0, the team-scoping logic could resolve a `conn_id` containing a `/` (e.g. `"my_team/conn"`) to the same path as another team's team-scoped secret when the caller had no team context. A privileged caller without team context could therefore retrieve another team's secret by crafting a colliding `conn_id`. Fixed in 9.28.0 by switching the team-scope separator to `--` and rejecting team-shaped `conn_id`s when team context is absent. Affects the experimental multi-tenant teams feature only. Users are recommended to upgrade to `a

• Auf 9.28.0 aktualisieren

Improper Origin Validation in mlflow/mlflow

In MLflow version 3.9.0, the MLflow Assistant feature introduced improper origin validation in its /ajax-api endpoints. This vulnerability allows a remote attacker to exploit cross-origin requests from a malicious webpage to interact with the MLflow Assistant running on a victim's local machine. By bypassing the loopback-only restriction, the attacker can modify the Assistant's configuration to enable full access, which in turn allows the execution of arbitrary commands via the Claude Code sub-agent. This issue is resolved in version 3.10.0.

• Auf 3.10.0 aktualisieren

CVE-2026-31072

The JSONSerializer and CBORSerializer in APScheduler (all versions including 3.10.x and 4.0.0a5) are vulnerable to Remote Code Execution (RCE) via Insecure Deserialization. The unmarshal_object function allows for arbitrary class instantiation and state injection by dynamically importing modules and calling __setstate__ on any class available in the Python environment. An attacker can exploit this by submitting a specially crafted JSON or CBOR payload to an application using these serializers

CVE-2025-51427

An issue was discovered in ModelScope 1.25.0 allowing attackers to execute arbitrary code via crafted module listed in the configuration file (dey_mini.yaml) under the key ['nnet']['module'].

• Auf 1.27.0 aktualisieren