NVIDIA Isaac-GR00T pour toutes les plateformes contient une vulnérabilité dans un composant Python, où un attaquant pourrait provoquer un problème d'injection de code. L'exploitation réussie de cette vulnérabilité pourrait mener à du code

The code injection vulnerability in NVIDIA Isaac-GR00T allows an attacker to inject and execute arbitrary code within the Python component. This could lead to a wide range of malicious activities, including gaining unauthorized access to sensitive data, modifying system configurations, and potentially taking control of the entire robotic system. The impact is particularly severe in environments where Isaac-GR00T is used for autonomous navigation or critical decision-making, as an attacker could manipulate the robot's behavior to cause harm or disruption. The ability to escalate privileges further amplifies the risk, allowing an attacker to move laterally within the system and compromise other connected resources. Data tampering could lead to inaccurate sensor readings or manipulated control signals, creating dangerous situations.

Organizations and individuals utilizing NVIDIA Isaac-GR00T for robotics development and deployment are at risk. This includes researchers, engineers, and companies developing autonomous systems, particularly those relying on custom Python scripts within their Isaac-GR00T workflows. Those using older, unpatched versions of Isaac-GR00T are most vulnerable.

• python / supply-chain:

import os
import subprocess

# Check for the vulnerable version of Isaac-GR00T
process = subprocess.Popen(['python', '-c', 'import isaac_gr00t; print(isaac_gr00t.__version__)'], stdout=subprocess.PIPE, stderr=subprocess.PIPE)
stdout, stderr = process.communicate()
version = stdout.decode('utf-8').strip()

if version and version != '7f53666':
    print(f"Vulnerable version detected: {version}")

• generic web: Check for unusual Python processes running with elevated privileges. • generic web: Monitor system logs for suspicious Python script executions or attempts to access sensitive files.

1. 2025-11-18Disclosure

   disclosure

1. Réservé2025-04-15
2. Publiée2025-11-18
3. Modifiée2025-11-19
4. EPSS mis à jour2026-03-23

The primary mitigation for CVE-2025-33184 is to upgrade NVIDIA Isaac-GR00T to version 7f53666 or later. If an immediate upgrade is not possible due to compatibility issues or system downtime constraints, consider implementing stricter input validation and sanitization within the Python component to prevent malicious code from being injected. While not a complete solution, this can reduce the attack surface. Monitor system logs for any unusual activity or attempts to execute unauthorized code. Implement robust access controls to limit who can modify the Python component and its dependencies. After upgrading, verify the fix by attempting to trigger the code injection vulnerability using known attack vectors and confirming that the attempts are blocked.