CVE-2025-33253: Vulnerability in nemo-toolkit Allows Remote Code Execution via Untrusted Data Deserialization
Summary
CVE-2025-33253: NVIDIA NeMo Framework Deserializes Untrusted Data. Affects nemo-toolkit versions < 2.6.1. Fixed in 2.6.1. How to fix ↓
Check if this vulnerability affects your application
Upload a manifest or SBOM and we'll show whether you're affected, plus every other risky package in the same scan.
No signup required · Results in about 60 seconds
Background
The NVIDIA NeMo Framework, specifically the nemo-toolkit package, is a powerful tool designed for building and training state-of-the-art conversational AI models. It provides a flexible and modular architecture that enables developers to create complex AI applications efficiently. However, the framework's reliance on deserializing data from external sources introduces significant security risks, particularly when handling untrusted data.
Vulnerability Detail
The vulnerability in the NVIDIA NeMo Framework arises from its ability to deserialize untrusted data, which can be exploited by an attacker to execute arbitrary code on the user's system. This issue can occur when a user is tricked into loading a maliciously crafted file, leading to potential remote code execution, denial of service, information disclosure, and data tampering. The vulnerability is particularly concerning due to its ease of exploitation and the severe consequences it can have on system integrity and confidentiality.
How To Fix
Latest published version: 3.0.0 as of 23 Oct 2026
pip install --upgrade "nemo-toolkit==3.0.0"Upgrading to it may include breaking changes; the version below is the minimum fixed release for this vulnerability.
If you need the minimum change
2.6.1
This is the first release that contains the fix.
pip install --upgrade "nemo-toolkit==2.6.1"Tip — Not sure which version your app runs? Vulert can match this fix to it. Scan free →
Workaround
As a temporary workaround, users should avoid loading untrusted files or data into the NeMo Framework until the package is updated. Implementing strict validation and sanitization of input data can also help reduce the risk of exploitation.
Conclusion
The vulnerability identified as CVE-2025-33253 in the NVIDIA NeMo Framework's nemo-toolkit package poses significant security risks due to its ability to deserialize untrusted data. Users are strongly advised to update to the patched version to safeguard their systems against potential exploits.
See every vulnerability recorded for nemo-toolkit →
FAQ
The CVSS score of this vulnerability is 7.8, indicating a high severity level.
You can use Vulert's playground at Vulert Scanner to check if your application is affected by this vulnerability without requiring any signup.
Exploitation of this vulnerability can lead to remote code execution, denial of service, information disclosure, and data tampering.
Yes, the vulnerability has been patched in version 2.6.1 of the nemo-toolkit package.
If immediate updates are not possible, avoid loading untrusted files and implement input validation to mitigate risks.
Related vulnerabilities
Recent vulnerabilities recorded for nemo-toolkit.
