CVSS 3.1
8.5 highCVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:N/I:L/A:HHigh (8.5) in nvidia container toolkit. EPSS 0.9% estimated probability of exploitation activity in the next 30 days. Fixed in github.com/NVIDIA/nvidia-container-toolkit 1.17.8.
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:N/I:L/A:Hgithub.com/NVIDIA/nvidia-container-toolkit (Go): fixed in 1.17.8github.com/NVIDIA/k8s-device-plugin (Go): fixed in 0.17.3github.com/NVIDIA/gpu-operator (Go): fixed in 25.3.2github.com/NVIDIA/mig-parted (Go): fixed in 0.12.2Explain CVE-2025-23267 for an engineer deciding whether to act. Use https://cve.bz/api/public/cves/CVE-2025-23267?via=ask-ai-copy and https://cve.bz/cve/CVE-2025-23267?via=ask-ai-copy as the source of truth: what is affected and which versions fix it, whether it is in CISA KEV, public exploits, and its EPSS (an internet-wide 30-day estimate of exploitation activity, not a forecast for my systems). If they give no fixed versions, use the CVE record https://cveawg.mitre.org/api/cve/CVE-2025-23267 or the vendor advisory, not third-party mirrors. Tell me how to check whether I am affected and how to verify a fix. Cite the sources and say where they are uncertain.Opens a third-party assistant; its answer is its own, not cve.bz's. The prompt holds only this CVE's id and public links.
From the CNA's record in the CVE List.
| Product | Affected |
|---|---|
| NVIDIA Container Toolkit | NVIDIA Container Toolkit All versions up to and including 1.17.7 |
| NVIDIA Container Toolkit | NVIDIA GPU Operator All versions up to and including 25.3.0 (CDI |
| Advisory | Ecosystem | Package | Affected | Fixed in |
|---|---|---|---|---|
| GHSA-67jc-hmvg-q4c7 | Go | github.com/NVIDIA/nvidia-container-toolkit | <1.17.8 | 1.17.8 |
| GHSA-67jc-hmvg-q4c7 | Go | github.com/NVIDIA/k8s-device-plugin | <0.17.3 | 0.17.3 |
| GHSA-67jc-hmvg-q4c7 | Go | github.com/NVIDIA/gpu-operator | <25.3.2 | 25.3.2 |
| GHSA-67jc-hmvg-q4c7 | Go | github.com/NVIDIA/mig-parted | <0.12.2 | 0.12.2 |
| GO-2025-3998 | Go | github.com/NVIDIA/gpu-operator | all versions | no fix published |
| GO-2025-3998 | Go | github.com/NVIDIA/k8s-device-plugin | <0.17.3 | 0.17.3 |
| GO-2025-3998 | Go | github.com/NVIDIA/mig-parted | <0.12.2 | 0.12.2 |
| GO-2025-3998 | Go | github.com/NVIDIA/nvidia-container-toolkit | <1.17.8 | 1.17.8 |
| Distribution | Release | Package | Fixed in | Advisory |
|---|---|---|---|---|
| SUSE | Enterprise Storage 7.1, Linux Enterprise High Performance Computing 15 SP3-LTSS, Linux Enterprise High Performance Computing 15 SP4-ESPOS and 11 more | nvidia-container-toolkit | 1.18.0-150200.5.17.1 | SUSE-SU-2025:4187-1 |
| SUSE | Linux Enterprise Module for Package Hub 15 SP6 | govulncheck-vulndb | 0.0.20251023T162509-150000.1.110.1 | SUSE-SU-2025:3799-1 |
| openSUSE | Leap 15.6 | govulncheck-vulndb | 0.0.20251023T162509-150000.1.110.1 | SUSE-SU-2025:3799-1 |
| openSUSE | Leap 15.6 | nvidia-container-toolkit | 1.18.0-150200.5.17.1 | SUSE-SU-2025:4187-1 |
| openSUSE | Leap 16.0 | govulncheck-vulndb | 0.0.20260723T184607-160000.1.1 | openSUSE-SU-2026:21483-1 |
| openSUSE | Tumbleweed | govulncheck-vulndb | 0.0.20251023T162509-1.1 | openSUSE-SU-2025:15666-1 |
From the distributions' data via OSV.dev: SUSE security data, CC BY 4.0. Modified: reformatted and linked to CVE records. Credits. On the distributions' own pages: SUSE · openSUSE.
Red Hat rates this Important.
RHEL AI is not affected because it uses CDI mode with NVIDIA Container Toolkit versions later than 1.17.5, which are not vulnerable, and does not rely on the enable-cuda-compat hook, thereby eliminating the attack vector entirely.
This vulnerability is rated as Important instead of Critical because, while it involves a symlink attack via the update-ldcache hook in the NVIDIA Container Toolkit, it has limited impact and specific requirements. The attacker needs low privileges inside a container and must use a specially crafted image to trigger the issue. It does not lead to code execution or privilege escalation, and it has no impact on confidentiality—only potential data tampering or denial of service. Additionally, the vulnerability affects only setups using this specific hook and can be easily mitigated by disabling it.
This vulnerability can be mitigated by opting out using the enable-cuda-compat hook. When using the NVIDIA Container Runtime in legacy mode, you can opt out of using the enable-cuda-compat hook by editing the /etc/nvidia-container-toolkit/config.toml file to set the features.disable-cuda-compat-lib-hook feature flag to true:
[features]
disable-cuda-compat-lib-hook = true
| Product | Package | State | Advisory |
|---|---|---|---|
| Red Hat Enterprise Linux 10 | toolbox | Not affected | |
| Red Hat Enterprise Linux 9 | toolbox | Not affected | |
| Red Hat Enterprise Linux AI (RHEL AI) | rhelai1/bootc-amd-rhel9 | Not affected | |
| Red Hat Enterprise Linux AI (RHEL AI) | rhelai1/bootc-aws-nvidia-rhel9 | Not affected | |
| Red Hat Enterprise Linux AI (RHEL AI) | rhelai1/bootc-azure-amd-rhel9 | Not affected | |
| Red Hat Enterprise Linux AI (RHEL AI) | rhelai1/bootc-azure-nvidia-rhel9 | Not affected | |
| Red Hat Enterprise Linux AI (RHEL AI) | rhelai1/bootc-gcp-nvidia-rhel9 | Not affected | |
| Red Hat Enterprise Linux AI (RHEL AI) | rhelai1/bootc-intel-rhel9 | Not affected | |
| Red Hat Enterprise Linux AI (RHEL AI) | rhelai1/bootc-nvidia-rhel9 | Not affected | |
| Red Hat Enterprise Linux AI (RHEL AI) | rhelai1/granite-3.1-8b-lab-v2.1 | Not affected | |
| Red Hat Enterprise Linux AI (RHEL AI) | rhelai1/granite-3.1-8b-starter-v2.1 | Not affected | |
| Red Hat Enterprise Linux AI (RHEL AI) | rhelai1/instructlab-amd-rhel9 | Not affected | |
| Red Hat Enterprise Linux AI (RHEL AI) | rhelai1/instructlab-nvidia-rhel9 | Not affected | |
| Red Hat Enterprise Linux AI (RHEL AI) | rhelai1/modelcar-granite-3-1-8b-lab-v2-1 | Not affected | |
| Red Hat Enterprise Linux AI (RHEL AI) | rhelai1/modelcar-granite-3-1-8b-starter-v2-1 | Not affected |
Contains Red Hat Security Data © Red Hat, Inc., licensed under CC BY 4.0. Source: https://access.redhat.com/security/data. Modified: reformatted. Red Hat's page for CVE-2025-23267.
The attack patterns MITRE's CAPEC relates to this record's weaknesses, as MITRE names them. Credits.