diffusers 代码注入漏洞

diffusers is a generative model library for generating images, audio, and 3D molecular structures, open-sourced by Hugging Face. Versions of diffusers prior to 0.38.0 contained a code injection vulnerability, which was caused by a bypass of trustremotecode, potentially allowing arbitrary remote...

GHSA-98H9-4798-4Q5V Diffusers has a `trust_remote_code` bypass via `custom_pipeline` and local custom components

Impact A trustremotecode bypass in DiffusionPipeline.frompretrained allows arbitrary remote code execution despite the user passing trustremotecode=False or omitting it, which is the default. The vulnerability has three variants, all sharing the same root cause — the trustremotecode gate was...

Diffusers has a `trust_remote_code` bypass via `custom_pipeline` and local custom components

Impact A trustremotecode bypass in DiffusionPipeline.frompretrained allows arbitrary remote code execution despite the user passing trustremotecode=False or omitting it, which is the default. The vulnerability has three variants, all sharing the same root cause — the trustremotecode gate was...

Arbitrary Code Injection

Overview diffusers is a State-of-the-art diffusion in PyTorch and JAX. Affected versions of this package are vulnerable to Arbitrary Code Injection in the frompretrained fucntion when a repository contains a None.py file and the custompipeline argument is not supplied. An attacker can execute...

Diffusers has a `trust_remote_code` bypass via `custom_pipeline` and local custom components

Background This vulnerability is found in the DiffusionPipeline.frompretrained flow, which is used to load a pipeline from the HuggingFace Hub. This function accepts an optional custompipeline keyword argument: the name of a Python file in the repo that contains a custom class inheriting from...

GHSA-J7W6-VPVQ-J3GM Diffusers has a `trust_remote_code` bypass via `custom_pipeline` and local custom components

Background This vulnerability is found in the DiffusionPipeline.frompretrained flow, which is used to load a pipeline from the HuggingFace Hub. This function accepts an optional custompipeline keyword argument: the name of a Python file in the repo that contains a custom class inheriting from...

PT-2026-39298

Name of the Vulnerable Software and Affected Versions Diffusers versions prior to 0.38.0 Description An issue exists in the DiffusionPipeline.from pretrained flow when loading pipelines from Hugging Face Hub repositories. The resolve custom pipeline and cls function in pipeline loading utils.py...

CVE-2026-7669

A vulnerability was detected in sgl-project SGLang up to 0.5.9. Impacted is the function gettokenizer of the file python/sglang/srt/utils/hftransformersutils.py of the component HuggingFace Transformer Handler. The manipulation of the argument trustremotecode with the input False as part of Boole...

InstructLab Includes Functionality from Untrusted Control Sphere

A flaw was found in InstructLab. The linuxtrain.py script hardcodes trustremotecode=True when loading models from HuggingFace. This allows a remote attacker to achieve arbitrary Python code execution by convincing a user to run ilab train/download/generate with a specially crafted malicious model...

EUVD-2026-24752

A flaw was found in InstructLab. The linuxtrain.py script hardcodes trustremotecode=True when loading models from HuggingFace. This allows a remote attacker to achieve arbitrary Python code execution by convincing a user to run ilab train/download/generate with a specially crafted malicious model...

CVE-2026-6859

A flaw was found in InstructLab. The linuxtrain.py script hardcodes trustremotecode=True when loading models from HuggingFace. This allows a remote attacker to achieve arbitrary Python code execution by convincing a user to run ilab train/download/generate with a specially crafted malicious model...

CVE-2026-6859 Instructlab: instructlab: arbitrary code execution due to hardcoded `trust_remote_code=true`

A flaw was found in InstructLab. The linuxtrain.py script hardcodes trustremotecode=True when loading models from HuggingFace. This allows a remote attacker to achieve arbitrary Python code execution by convincing a user to run ilab train/download/generate with a specially crafted malicious model...

CVE-2026-6859 Instructlab: instructlab: arbitrary code execution due to hardcoded `trust_remote_code=true`

A flaw was found in InstructLab. The linuxtrain.py script hardcodes trustremotecode=True when loading models from HuggingFace. This allows a remote attacker to achieve arbitrary Python code execution by convincing a user to run ilab train/download/generate with a specially crafted malicious model...

PT-2026-34336

Name of the Vulnerable Software and Affected Versions InstructLab affected versions not specified Description A flaw exists in the linux train.py script which hardcodes the trust remote code variable as True when loading models from HuggingFace. This allows a remote attacker to achieve arbitrary...

Red Hat Enterprise Linux AI 安全漏洞

Red Hat Enterprise Linux AI is a Linux distribution created by the American company Red Hat for generative AI. Red Hat Enterprise Linux AI RHEL AI 3 has a security vulnerability. This vulnerability stems from the linuxtrain.py script, which loads models from HuggingFace by hardcoding...

vLLM has Hardcoded Trust Override in Model Files Enables RCE Despite Explicit User Opt-Out

Summary Two model implementation files hardcode trustremotecode=True when loading sub-components, bypassing the user's explicit --trust-remote-code=False security opt-out. This enables remote code execution via malicious model repositories even when the user has explicitly disabled remote code...

Unsafe Dependency Resolution

Overview vllm is an A high-throughput and memory-efficient inference and serving engine for LLMs Affected versions of this package are vulnerable to Unsafe Dependency Resolution in the process of loading sub-components with the trustremotecode parameter set to True, regardless of user...

CVE-2026-27893

vLLM is an inference and serving engine for large language models LLMs. Starting in version 0.10.1 and prior to version 0.18.0, two model implementation files hardcode trustremotecode=True when loading sub-components, bypassing the user's explicit --trust-remote-code=False security opt-out. This...

PYSEC-2026-2297

vLLM is an inference and serving engine for large language models LLMs. Starting in version 0.10.1 and prior to version 0.18.0, two model implementation files hardcode trustremotecode=True when loading sub-components, bypassing the user's explicit --trust-remote-code=False security opt-out. This...

CVE-2026-27893 vLLM's hardcoded trust_remote_code=True in NemotronVL and KimiK25 bypasses user security opt-out

vLLM is an inference and serving engine for large language models LLMs. Starting in version 0.10.1 and prior to version 0.18.0, two model implementation files hardcode trustremotecode=True when loading sub-components, bypassing the user's explicit --trust-remote-code=False security opt-out. This...