Skip to main content
CVE Vulnerability Database
Vulnerability Database/CVE-2025-23347

CVE-2025-23347: NVIDIA G-Assist Privilege Escalation Flaw

CVE-2025-23347 is a privilege escalation vulnerability in NVIDIA Project G-Assist that enables attackers to elevate permissions and execute code. This article covers technical details, affected versions, and mitigation.

Published:

CVE-2025-23347 Overview

CVE-2025-23347 is a local privilege escalation vulnerability in NVIDIA Project G-Assist. An authenticated local attacker can leverage improper permission assignment [CWE-276] to escalate privileges on affected systems. Successful exploitation can lead to code execution, escalation of privileges, data tampering, denial of service, and information disclosure.

The vulnerability requires local access and low privileges but no user interaction. NVIDIA published guidance for this issue in NVIDIA Support Answer 5703.

Critical Impact

An authenticated local user can escalate to higher privileges and gain full confidentiality, integrity, and availability impact on the host running NVIDIA Project G-Assist.

Affected Products

  • NVIDIA Project G-Assist
  • Windows systems with NVIDIA Project G-Assist installed
  • Refer to the NVIDIA Support Answer for specific affected versions

Discovery Timeline

  • 2025-10-23 - CVE-2025-23347 published to NVD
  • 2026-06-17 - Last updated in NVD database

Technical Details for CVE-2025-23347

Vulnerability Analysis

CVE-2025-23347 is classified as an improper permission assignment for a critical resource [CWE-276]. NVIDIA Project G-Assist contains a resource whose access controls do not adequately restrict lower-privileged users. A local attacker with existing low-privilege access can interact with that resource to gain elevated permissions.

Because the vulnerability supports full impact on confidentiality, integrity, and availability, exploitation can result in code execution in a higher security context. This enables data tampering, disruption of services, and disclosure of sensitive information.

Root Cause

The root cause is improper permission assignment on a resource shipped or configured by NVIDIA Project G-Assist. When a securable object such as a file, directory, registry key, service, or named pipe grants write or control access to non-administrative users, that user can replace or influence a component that later executes with elevated rights. NVIDIA's advisory does not disclose the specific resource, and the NVD entry for CVE-2025-23347 references NVIDIA's bulletin for full technical detail.

Attack Vector

The attack vector is local. An attacker must already have valid credentials and code execution on the target host as a standard user. From that foothold, the attacker manipulates the mis-permissioned resource used by NVIDIA Project G-Assist to trigger execution under a higher-privileged process. No user interaction is required, and the scope remains unchanged, meaning the elevated context is on the same security authority as the vulnerable component.

No public proof-of-concept exploit is currently listed, and the vulnerability is not on the CISA Known Exploited Vulnerabilities catalog.

Detection Methods for CVE-2025-23347

Indicators of Compromise

  • Unexpected modification of NVIDIA Project G-Assist installation files, directories, or registry keys by non-administrative accounts.
  • Child processes spawned by NVIDIA G-Assist services or scheduled tasks that do not match expected binaries or signing.
  • New or altered DLLs, executables, or configuration files inside G-Assist directories that were not written by an NVIDIA installer.

Detection Strategies

  • Audit the access control lists on NVIDIA Project G-Assist installation directories, services, and named pipes for entries granting write or modify rights to Users, Authenticated Users, or Everyone.
  • Monitor process creation events where a G-Assist service or helper process launches an unsigned or unexpected child binary.
  • Correlate file write events in G-Assist directories with the security context of the writing process to identify low-privileged tampering.

Monitoring Recommendations

  • Enable Windows Security event logging for object access on NVIDIA Project G-Assist files and registry keys, then forward to a SIEM for review.
  • Alert on service configuration changes (sc config) or scheduled task modifications targeting NVIDIA components.
  • Track integrity of NVIDIA Project G-Assist binaries with file integrity monitoring across managed endpoints.

How to Mitigate CVE-2025-23347

Immediate Actions Required

  • Apply the fixed version of NVIDIA Project G-Assist as described in NVIDIA Support Answer 5703.
  • Inventory all endpoints and workstations running NVIDIA Project G-Assist and prioritize systems used by privileged users.
  • Restrict interactive logon on hosts running G-Assist to trusted users while the patch is being deployed.

Patch Information

NVIDIA has published remediation guidance in its security bulletin. Refer to NVIDIA Support Answer 5703 for the fixed version and download instructions. Additional references are available in the NVD record and the CVE.org record for CVE-2025-23347.

Workarounds

  • Uninstall NVIDIA Project G-Assist on systems that do not require it until the update can be applied.
  • Tighten NTFS and registry ACLs on NVIDIA Project G-Assist resources to remove write permissions for non-administrative users, following vendor guidance.
  • Apply the principle of least privilege so local users cannot run arbitrary code on hosts where G-Assist is installed.
bash
# Configuration example - review permissions on the G-Assist install path
icacls "C:\Program Files\NVIDIA Corporation\G-Assist"
# Remove write access for non-admin groups if present (adjust path per vendor guidance)
icacls "C:\Program Files\NVIDIA Corporation\G-Assist" /remove:g "Users" "Authenticated Users"

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Default Legacy - Prefooter | Experience the World’s Most Advanced Cybersecurity Platform

Experience the Most Advanced Cybersecurity Platform

See how the world’s most intelligent, autonomous cybersecurity platform can protect your organization today and into the future.