Security vulnerabilities, privacy incidents, safety concerns, and policy updates affecting LLMs and AI agents.
PyTorch (a Python package for tensor computation) versions before 2.10.0 have a vulnerability in the `weights_only` unpickler that allows attackers to create malicious checkpoint files (.pth files, which store model data) triggering memory corruption and potentially arbitrary code execution (running attacker-chosen commands) when loaded with `torch.load(..., weights_only=True)`. This is a deserialization vulnerability (a weakness where loading untrusted data can be exploited).
Fix: Update to PyTorch version 2.10.0 or later, which fixes the issue.
NVD/CVE DatabaseCVE-2025-55558 is a buffer overflow (a memory safety error where data is written beyond the intended boundaries) in PyTorch version 2.7.0 that occurs when certain neural network operations are combined and compiled using Inductor, a code compiler. This vulnerability causes a Denial of Service attack (making a service unavailable to users), though no CVSS severity score has been assigned yet.
CVE-2024-6577 is a vulnerability in PyTorch Serve where a script called 'upload_results_to_s3.sh' references an Amazon S3 bucket (a cloud storage service) without verifying that the script's creators actually own or control it, potentially allowing unauthorized access to sensitive data stored in that bucket.
PyTorch versions before 2.2.0 contain an out-of-bounds read vulnerability (a bug where code tries to read data from memory outside its allowed range) in the flatbuffer_loader component, which is used for loading machine learning models on mobile devices. This vulnerability could potentially allow attackers to read sensitive information from memory or cause the program to crash.
PyTorch versions before v2.2.0 contain a use-after-free vulnerability (a memory bug where code tries to access data that has already been freed) in the mobile interpreter component. This vulnerability was identified in the torch/csrc/jit/mobile/interpreter.cpp file.
Fix: Upgrade to PyTorch version 2.2.0 or later. A patch is available at https://github.com/pytorch/pytorch/commit/7c35874ad664e74c8e4252d67521f3986eadb0e6.
NVD/CVE DatabaseFix: Update PyTorch to version v2.2.0 or later. A patch is available at https://github.com/pytorch/pytorch/commit/9c7071b0e324f9fb68ab881283d6b8d388a4bcd2.
NVD/CVE Database