Supply chain · pip · High

Malicious pip package 'nvtorch-oot-nightly' exfiltrates host information

Threat Engine Auto-Feed · data current as of 2026-08-01

The pip package `nvtorch-oot-nightly` version `99999.0.0` contains malicious code designed to exfiltrate basic host information upon installation or import. This package has no legitimate purpose beyond this activity, posing a critical supply-chain risk for Python environments.

  • Package: pip / nvtorch-oot-nightly
  • Affected versions: = 99999.0.0
  • Severity: critical
  • Reach: unknown
  • Exposure: 65/100 (High)

What happened — Credential / secret theft

The `nvtorch-oot-nightly` package on PyPI, specifically version `99999.0.0`, was found to contain malicious code. This code is executed either during the package's installation process or when the module is imported, leading to the exfiltration of basic host information. The package's `setup.py` overrides the install command to facilitate this malicious execution.

How the attack works

This is a software supply-chain attack where a malicious package is published to a public registry. When a user or automated system installs or imports the package, it triggers the execution of embedded malicious code. This code then collects and transmits basic system data, such as IP address or username, to an external location controlled by the attacker. The attack leverages the package's installation hooks or module import mechanisms to achieve execution.

Who's exposed

Profiles most at risk

  • Development teams using Python and pip
  • CI/CD pipelines that automatically pull Python dependencies
  • Organizations with internal mirrors that sync with PyPI

Conditions that increase exposure

  • Projects with unpinned dependencies that might inadvertently pull the malicious version
  • Environments lacking `pip freeze` or `requirements.txt` lockfiles
  • Automated build systems configured to install the latest available package versions

Blast-radius scenarios

  • An attacker gains access to basic host information (e.g., IP, username) from any system where the package is installed or imported.
  • Compromised build environments could lead to further lateral movement or exposure of build secrets if the exfiltrated data aids in subsequent attacks.

What to do (defensive)

Detect

  • Scan `requirements.txt` and `setup.py` files for `nvtorch-oot-nightly` version `99999.0.0`.
  • Monitor network traffic for suspicious outbound connections from build systems or development environments that have installed Python packages.
  • Review package installation logs for unusual commands executed by `setup.py`.

Contain

  • Immediately quarantine any systems or environments found to have installed `nvtorch-oot-nightly` version `99999.0.0`.
  • Block `nvtorch-oot-nightly` version `99999.0.0` from being downloaded from PyPI or internal package mirrors.
  • Isolate development and build environments from production networks until verified clean.

Remediate

  • Remove `nvtorch-oot-nightly` from all project dependencies.
  • Rebuild affected environments from trusted sources, ensuring all dependencies are explicitly pinned to known good versions.
  • Rotate any credentials or API keys that may have been present on compromised systems, as basic host information could aid in targeted credential theft.

Frontier verdict — High

Prioritize immediate detection and removal of `nvtorch-oot-nightly` due to its critical severity and data exfiltration capabilities.

Sources

For detection-engineering and awareness only · point-in-time · not security advice · sourced from the GitHub Advisory Database.