Логотип exploitDog
Консоль
Логотип exploitDog

exploitDog

redhat логотип

CVE-2026-31221

Опубликовано: 12 мая 2026
Источник: redhat
CVSS3: 8
EPSS Низкий

Описание

PyTorch-Lightning versions 2.6.0 and earlier contain an insecure deserialization vulnerability (CWE-502) in the checkpoint loading mechanism. The LightningModule.load_from_checkpoint() method, which is commonly used to load saved model states, internally calls torch.load() without setting the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the Pickle module. A remote attacker can exploit this by providing a maliciously crafted checkpoint file, leading to arbitrary code execution on the victim's system when the file is loaded.

A flaw was found in PyTorch-Lightning. This vulnerability, categorized as insecure deserialization (CWE-502), exists in the checkpoint loading mechanism. A remote attacker can exploit this by providing a maliciously crafted checkpoint file, which, when loaded by the LightningModule.load_from_checkpoint() method, allows the deserialization of arbitrary Python objects. This can lead to arbitrary code execution on the victim's system.

Отчет

This is an Important arbitrary code execution flaw in PyTorch-Lightning due to insecure deserialization in its checkpoint loading mechanism. Red Hat products utilizing PyTorch-Lightning, such as Red Hat Enterprise Linux AI and Red Hat AI Inference Server, are affected if they process untrusted checkpoint files. An attacker can craft a malicious checkpoint file, which, when loaded, can lead to arbitrary code execution on the system.

Меры по смягчению последствий

To mitigate this issue, avoid loading PyTorch-Lightning checkpoint files from untrusted or unverified sources. Ensure that all checkpoint files originate from trusted environments to prevent the deserialization of malicious Python objects and subsequent arbitrary code execution.

Затронутые пакеты

ПлатформаПакетСостояниеРекомендацияРелиз
Red Hat AI Inference Serverrhaiis/vllm-cpu-rhel9Will not fix
Red Hat AI Inference Serverrhaiis/vllm-cuda-rhel9Affected
Red Hat AI Inference Serverrhaiis/vllm-rocm-rhel9Affected
Red Hat AI Inference Serverrhaiis/vllm-tpu-rhel9Will not fix
Red Hat Enterprise Linux AI (RHEL AI) 3rhelai3/bootc-aws-cuda-rhel9Affected
Red Hat Enterprise Linux AI (RHEL AI) 3rhelai3/bootc-azure-cuda-rhel9Affected
Red Hat Enterprise Linux AI (RHEL AI) 3rhelai3/bootc-azure-rocm-rhel9Affected
Red Hat Enterprise Linux AI (RHEL AI) 3rhelai3/bootc-cuda-rhel9Affected
Red Hat Enterprise Linux AI (RHEL AI) 3rhelai3/bootc-gcp-cuda-rhel9Affected
Red Hat Enterprise Linux AI (RHEL AI) 3rhelai3/bootc-rocm-rhel9Affected

Показывать по

Дополнительная информация

Статус:

Important
Дефект:
CWE-502
https://bugzilla.redhat.com/show_bug.cgi?id=2476502pytorch-lightning: PyTorch-Lightning: Arbitrary code execution via insecure deserialization of checkpoint files

EPSS

Процентиль: 31%
0.00385
Низкий

8 High

CVSS3

Связанные уязвимости

CVSS3: 7.8
ubuntu
3 месяца назад

PyTorch-Lightning versions 2.6.0 and earlier contain an insecure deserialization vulnerability (CWE-502) in the checkpoint loading mechanism. The LightningModule.load_from_checkpoint() method, which is commonly used to load saved model states, internally calls torch.load() without setting the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the Pickle module. A remote attacker can exploit this by providing a maliciously crafted checkpoint file, leading to arbitrary code execution on the victim's system when the file is loaded.

CVSS3: 7.8
nvd
3 месяца назад

PyTorch-Lightning versions 2.6.0 and earlier contain an insecure deserialization vulnerability (CWE-502) in the checkpoint loading mechanism. The LightningModule.load_from_checkpoint() method, which is commonly used to load saved model states, internally calls torch.load() without setting the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the Pickle module. A remote attacker can exploit this by providing a maliciously crafted checkpoint file, leading to arbitrary code execution on the victim's system when the file is loaded.

CVSS3: 7.8
github
3 месяца назад

PyTorch Lightning load_from_checkpoint has an insecure checkpoint deserialization

EPSS

Процентиль: 31%
0.00385
Низкий

8 High

CVSS3