PT-2026-27512

Name of the Vulnerable Software and Affected Versions NVIDIA NeMo Framework affected versions not specified Description The NVIDIA NeMo Framework has a flaw in how it loads checkpoints, potentially allowing a remote attacker to execute code. Exploitation of this issue could result in code...

NVIDIA Nemo Framework code-related vulnerabilities

NVIDIA Nemo Framework is a framework developed by NVIDIA Corporation in the United States for building and deploying generative AI models. The NVIDIA NeMo Framework has code-related vulnerabilities, which stem from checkpoint loading issues. These vulnerabilities may lead to remote code execution...

Security Bulletin: NVIDIA Megatron LM - March 2026

NVIDIA has released a software update for NVIDIA® Megatron LM. To protect your system, clone or update this software to version 0.15.3 or later from NVIDIA/Megatron-LM on GitHub. Go to NVIDIA Product Security. Details The following table summarizes the potential vulnerabilities that this security...

CVE-2026-24152: Deserialization of Untrusted Data

NVIDIA Megatron-LM contains a vulnerability in checkpoint loading where an Attacker may cause an RCE by convincing a user to load a maliciously crafted file. A successful exploit of this vulnerability may lead to code execution, escalation of privileges, information disclosure, and data tampering...

NVIDIA Megatron-LM code-related vulnerabilities

NVIDIA Megatron-LM is a distributed training framework based on PyTorch developed by NVIDIA Corporation in the United States. It is specifically designed for training large-scale Transformer language models. NVIDIA Megatron-LM has code vulnerabilities, which stem from checkpoint loading issues...

PT-2026-27509

NVIDIA Megatron-LM contains a vulnerability in checkpoint loading where an Attacker may cause an RCE by convincing a user to load a maliciously crafted file. A successful exploit of this vulnerability may lead to code execution, escalation of privileges, information disclosure, and data tampering...

NVIDIA Megatron-LM code-related vulnerabilities

NVIDIA Megatron-LM is a distributed training framework based on PyTorch developed by NVIDIA Corporation in the United States. It is specifically designed for training large-scale Transformer language models. NVIDIA Megatron-LM has code vulnerabilities, particularly an issue related to the...

CVE-2026-24150: Deserialization of Untrusted Data

NVIDIA Megatron-LM contains a vulnerability in checkpoint loading where an Attacker may cause an RCE by convincing a user to load a maliciously crafted file. A successful exploit of this vulnerability may lead to code execution, escalation of privileges, information disclosure, and data tampering...

CVE-2026-24157: Deserialization of Untrusted Data

NVIDIA NeMo Framework contains a vulnerability in checkpoint loading where an attacker could cause remote code execution. A successful exploit of this vulnerability might lead to code execution, escalation of privileges, information disclosure and data tampering...

CVE-2026-28277 LangGraph: Unsafe msgpack deserialization in LangGraph checkpoint loading

LangGraph SQLite Checkpoint is an implementation of LangGraph CheckpointSaver that uses SQLite DB both sync and async, via aiosqlite. In version 1.0.9 and prior, LangGraph checkpointers can load msgpack-encoded checkpoints that reconstruct Python objects during deserialization. If an attacker can...

CVE-2026-28277 LangGraph: Unsafe msgpack deserialization in LangGraph checkpoint loading

LangGraph SQLite Checkpoint is an implementation of LangGraph CheckpointSaver that uses SQLite DB both sync and async, via aiosqlite. In version 1.0.9 and prior, LangGraph checkpointers can load msgpack-encoded checkpoints that reconstruct Python objects during deserialization. If an attacker can...

CVE-2026-28277 LangGraph: Unsafe msgpack deserialization in LangGraph checkpoint loading

LangGraph SQLite Checkpoint is an implementation of LangGraph CheckpointSaver that uses SQLite DB both sync and async, via aiosqlite. In version 1.0.9 and prior, LangGraph checkpointers can load msgpack-encoded checkpoints that reconstruct Python objects during deserialization. If an attacker can...

CVE-2026-28277

CVE-2026-28277 affects LangGraph checkpoint loading using an SQLite-backed checkpoint store. In versions ≤1.0.9, msgpack-encoded checkpoints may deserialize into Python objects, enabling an attacker with write access to the checkpoint store to craft payloads that trigger unsafe reconstruction whe...

Arbitrary Code Execution via Unsafe torch.load() in Trainer Checkpoint Loading

Summary A critical arbitrary code execution vulnerability exists in HuggingFace Transformers' Trainer class. The loadrngstate method at src/transformers/trainer.py:3059 calls torch.load without the weightsonly=True parameter. While a safeglobals context manager wraps this call, it provides no...

CVE-2025-67729

LMDeploy prior to v0.11.1 is affected by an insecure deserialization vulnerability in torch.load() called without weights_only=True when loading model checkpoint files (.bin/.pt). This can allow an attacker to execute arbitrary code on the victim's machine. The issue is patched in v0.11.1. Affect...

CVE-2025-67729 lmdeploy vulnerable to Arbitrary Code Execution via Insecure Deserialization in torch.load()

LMDeploy is a toolkit for compressing, deploying, and serving LLMs. Prior to version 0.11.1, an insecure deserialization vulnerability exists in lmdeploy where torch.load is called without the weightsonly=True parameter when loading model checkpoint files. This allows an attacker to execute...

lmdeploy code-related vulnerabilities

lmdeploy is an InternLM open source toolkit for compressing, deploying and servicing LLM. A code issue vulnerability exists in versions of lmdeploy prior to 0.11.1 that stems from loading model checkpoint files without using the weightsonly parameter, which could lead to an attacker executing...

PT-2025-48583

Name of the Vulnerable Software and Affected Versions Tencent NeuralNLP-NeuralClassifier affected versions not specified Description A flaw exists within the load checkpoint function that allows remote attackers to execute arbitrary code on affected installations. The issue stems from insufficien...

EUVD-2025-27190

Malicious code in bioql PyPI...

CVE-2025-58756

MONAI Medical Open Network for AI is an AI toolkit for health care imaging. In versions up to and including 1.5.0, in modeldict = torch.loadfullpath, maplocation=torch.devicedevice, weightsonly=True in monai/bundle/scripts.py , weightsonly=True is loaded securely. However, insecure loading method...