What is the CVSS score of CVE-2026-24155?

CVE-2026-24155 has a CVSS 3.1 base score of 7.8 (High). CVSS vector: CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H. EPSS exploitation probability: 0.2%.

Which versions of NVIDIA NeMo Framework are affected by CVE-2026-24155?

A code injection vulnerability (CVE-2026-24155) in NVIDIA NeMo Framework enables local attackers with low privileges to execute arbitrary code and escalate privileges, directly threatening the…

How to fix or mitigate CVE-2026-24155?

Within 24 hours: Inventory all NVIDIA NeMo Framework deployments across production and development environments; categorize by data sensitivity and assess current privilege levels of service accounts running NeMo. Within 7 days: Implement OS-level access restrictions confining NeMo process execution to minimal required privileges; deploy host-based detection (EDR) to monitor process spawning and privilege escalation from NeMo systems; review authentication and privilege escalation logs for the past 90 days. Within 30 days: Establish direct escalation with NVIDIA to obtain patch timeline and interim guidance; evaluate migration to alternative ML frameworks or containerized/sandboxed deployment models if vendor patch timeline exceeds organizational risk tolerance; deploy network segmentation isolating NeMo systems from authentication infrastructure and sensitive data repositories.

NVIDIA NeMo Framework CVE-2026-24155

EUVD-2026-37129 HIGH

Code Injection (CWE-94)

2026-06-16 nvidia

GHSA-xp23-6mrm-wqh3

Nvidia Code Injection Information Disclosure RCE Nemo Framework

7.8

CVSS 3.1 · Vendor: nvidia

Severity by source

Vendor (nvidia) PRIMARY

7.8 HIGH

AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H

vuln.today AI

7.8 HIGH

Local attacker with low-privileged shell on a NeMo host supplies a crafted artifact NeMo evaluates, yielding code execution with full CIA impact under the NeMo process; no UI or scope change.

3.1 AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H

4.0 AV:L/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N

Primary rating from Vendor (nvidia).

CVSS VectorVendor: nvidia

CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H

Attack Vector

Local

Attack Complexity

Low

Privileges Required

Low

User Interaction

None

Scope

Unchanged

Confidentiality

High

Integrity

High

Availability

High

Lifecycle Timeline

Analysis Generated

Jun 16, 2026 - 17:17 vuln.today

DescriptionCVE.org

NVIDIA NeMo Framework for all platforms contains a code injection vulnerability. A successful exploit of this vulnerability might lead to code execution, escalation of privileges, information disclosure, and data tampering.

AnalysisAI

Code injection in NVIDIA NeMo Framework across all supported platforms allows a local attacker with low privileges to execute arbitrary code, escalate privileges, disclose sensitive information, and tamper with data. The flaw carries a CVSS 3.1 score of 7.8 with high impact across confidentiality, integrity, and availability, though no public exploit identified at time of analysis and the issue is not listed in CISA KEV.

Technical ContextAI

NVIDIA NeMo Framework is an end-to-end, cloud-native toolkit used by researchers and enterprises to build, customize, and deploy generative AI and large language models, often executed in Python-based training and inference pipelines on GPU-equipped hosts. The CPE 'cpe:2.3:a:nvidia:nemo_framework:*' indicates all versions are in scope pending vendor clarification. The root cause is CWE-94 (Improper Control of Generation of Code, 'Code Injection'), meaning attacker-controlled input is incorporated into something the framework subsequently evaluates or executes - typical patterns in ML frameworks include unsafe deserialization of model artifacts, dynamic import/eval of configuration files (e.g., Hydra/OmegaConf, YAML, pickle), or template expansion in training recipes.

RemediationAI

Patch available per vendor advisory - consult the NVIDIA security bulletin at https://nvidia.custhelp.com/app/answers/detail/a_id/5839 for the exact fixed NeMo Framework version and upgrade all training, fine-tuning, and inference environments accordingly; the NVD record at https://nvd.nist.gov/vuln/detail/CVE-2026-24155 should be monitored for additional CPE/version metadata. Until upgraded, treat NeMo as an untrusted code execution surface: restrict who can submit configs, recipes, and model artifacts to NeMo pipelines, run NeMo workloads under dedicated non-privileged service accounts with minimal filesystem and network egress, and avoid loading model checkpoints, YAML/Hydra configs, or pickled artifacts from untrusted sources (the trade-off is operational friction for data scientists who routinely pull community models from hubs). Adding mandatory provenance and integrity checks (signatures or hash allowlists) on ingested model and config files reduces exposure but requires pipeline changes.