CVE-2026-7669-PoC

CVE-2026-7669 SGLang silently overrides trustremotecode=False to True and re-invokes AutoTokenizer.frompretrained whenever transformers v5 returns a TokenizersBackend object. A model with a custom tokenizerclass and an automap pointing at tokenizer.py reaches arbitrary code execution inside the...

CVE-2024-11393

CVE-2024-11393 Hugging Face Transformers MaskFormer Model Deserialization of Untrusted Data Remote Code Execution Vulnerability High-level overview and effects of the vulnerability: The vulnerability allows an attacker to execute arbitrary code on the host machine by supplying a malicious...

attack-attention

Attacking Attention of Foundation Models Effectively Disrupts Downstream Tasks Official PyTorch implementation of the paper "Attacking Attention of Foundation Models Effectively Disrupts Downstream Tasks" , accepted in the Adversarial Machine Learning on Computer Vision: Foundation Models + X ADV...

CVE-2026-22807_Range

🛡️ CVE-2026-22807 Simulation: AI Supply Chain RCE Target Range ⚠️ Disclaimer This project is intended solely for security research and educational purposes , aiming to demonstrate the supply chain attack principle Remote Code Execution during AI model loading. Do not use the generated malicious...

Noctyra

Noctyra Noctyra is an AST-based framework designed for code transformation and deobfuscation. It provides a modular pipeline to analyze and simplify Python source code by resolving complex expressions and logic. Overview The project aims to provide a flexible environment for processing Python's...

EUVD-2026-107004

vLLM: Request-controlled mmprocessorkwargs.coderevision allows remote code execution...

CVE-2026-23001-Hugging-Face-Transformers-Model-Deserialization-Arbitrary-Code-via-Pickle-

CVE-2026-23001 – Hugging Face Transformers Model Deserialization RCE Overview Hugging Face models saved with torch.save use Python’s pickle serialization, which is inherently unsafe. If a user downloads a seemingly legitimate model from a repository, loading it can trigger arbitrary code executio...

llama-factory-webui-rce-cve-2026-58116-trust-remote-code-model-path-injection

CVE-2026-58116 — LLaMA-Factory WebUI RCE via trustremotecode Proof of concept for CVE-2026-58116 CVSS 9.8 Critical: remote code execution in LLaMA-Factory = v0.9.5. The LLaMA-Factory WebUI hardcodes trustremotecode=True when loading a model src/llamafactory/webui/chatter.py:139 and...

CVE-2024-11392

CVE-2024-11392 Hugging Face Transformers MobileViTV2 Deserialization of Untrusted Data Remote Code Execution Vulnerability High-level overview and effects of the vulnerability: The vulnerability allows an attacker to execute arbitrary code on the host machine by supplying a malicious configuratio...

CVE-2024-11394

CVE-2024-11394 Hugging Face Transformers Trax Model Deserialization of Untrusted Data Remote Code Execution Vulnerability High-level overview and effects of the vulnerability: The vulnerability allows an attacker to execute arbitrary code on the host machine by supplying a malicious reformer trax...

Pickle-Deserialization-Exploit-in-Transformers

Exploiting Pickle Deserialization Vulnerability in Transformers Experiment Introduction Introduces the pickle deserialization mechanism and its reduce attack principle, reproduces the deserialization vulnerability CVE-2024-3568 in the Hugging Face Transformers library, analyzes security risks in...

Duplicate Advisory: vLLM: Request-controlled mm_processor_kwargs.code_revision allows remote code execution

Duplicate Advisory This advisory has been withdrawn because it is a duplicate of GHSA-h3rc-6mm3-gc2m. This link is maintained to preserve external references. Original Description vLLM from 0.7.3 before 0.31.0 forwards the request-supplied mmprocessorkwargs.coderevision to Transformers...

PT-2026-120345

A weakness has been identified in Tenstorrent tt-metal up to 0.80.0. The affected element is the function load hf state dict filtered of the file models/tt transformers/tt/load checkpoints.py. This manipulation of the argument weight map causes path traversal. The attack may be initiated remotely...

PT-2026-120179

Name of the Vulnerable Software and Affected Versions vLLM versions 0.7.3 through 0.30.x Description When trust remote code is enabled, the software forwards the request-supplied mm processor kwargs.code revision to the Transformers AutoProcessor.from pretrained function. This allows clients to...

vLLM code injection vulnerability

Versions vLLM 0.7.3 through 0.31.0 contain a code injection vulnerability. This vulnerability arises when the mmprocessorkwargs.coderevision parameter provided by the request is passed to Transformers AutoProcessor.frompretrained. When trustremotecode is enabled, it allows clients to specify whic...

Model Exposure on One Device Endangers Others: Rethinking Edge Neural Network Security across Devices

Neural networks NNs are increasingly deployed on edge devices such as mobile and IoT platforms, offering low latency and improved privacy compared to cloud-based deployment. However, on-device models embedded in mobile applications can be unpacked and exposed to adversaries, enabling cross-node...

Security Bulletin: Maximo AI Service uses multiple third party dependencies which are vulnerable to multiple CVEs.

Summary Maximo AI Service uses transformers-5.8.0-py3-none-any.whl, mlflow-3.11.1-py3-none-any.whl, sqlparse-0.5.4-py3-none-any.whl, nltk-3.10.0-py3-none-any.whl, gitpython-3.1.53-py3-none-any.whl, and cryptography-46.0.7-cp311-abi3-macosx109universal2.whl which are vulnerable to CVE-2026-80047,...

CVE-2026-80047

A flaw was found in the Hugging Face Transformers library. A remote attacker could exploit this vulnerability by enticing a user to load a model using the GenerativePreTrainedModel.loadcustomgenerate function. This function writes remote Python files to the local disk without requiring explicit...

Hugging Face Transformers downloads custom generation code before trust consent

A vulnerability in Hugging Face Transformers versions 4.49.0, = 5.8.1 allows remote Python files to be written to local disk without user consent when using GenerativePreTrainedModel.loadcustomgenerate. The function fetches and caches a remote module file before performing the required...

PYSEC-2026-4174 Hugging Face Transformers downloads custom generation code before trust consent

A vulnerability in Hugging Face Transformers versions 4.49.0, = 5.8.1 allows remote Python files to be written to local disk without user consent when using GenerativePreTrainedModel.loadcustomgenerate. The function fetches and caches a remote module file before performing the required...