CVE-2025-64750
Linux Ubuntu vulnerability analysis and mitigation

Overview

CVE-2025-64750 is an LSM (Linux Security Module) label bypass vulnerability in SingularityCE and SingularityPRO container platforms, classified as "Ineffective application of SELinux/AppArmor LSM process labels." It affects SingularityCE versions prior to 4.3.5 and SingularityPRO versions prior to 4.1.11 and 4.3.5. The vulnerability was disclosed on December 2, 2025, with patches released simultaneously. It carries a CVSS v3.1 base score of 4.5 (Medium) (GitHub Advisory).

Technical details

The root cause is improper resolution of the /proc filesystem path when writing LSM process labels (CWE-61: UNIX Symbolic Link Following; CWE-706: Use of Incorrectly-Resolved Name or Reference). When a user applies AppArmor or SELinux restrictions via the --security flag, Singularity writes labels to process/thread attr/exec entries under /proc. A malicious container image can redirect the mount of /proc to a shared mount destination, causing the LSM label write to target a non-procfs file and silently fail — leaving the container process without the intended security restrictions. Additionally, a bug in SELinux support detection caused --security selinux: flags to silently emit a warning rather than a fatal error when SELinux was unavailable, potentially going unnoticed in scripted environments. The fix involved switching to the pathrs-lite/procfs API to ensure writes target a genuine procfs attr/exec, and converting silent warnings to fatal errors (GitHub Advisory, Commit 5af3e79, Commit 2788296).

Impact

Successful exploitation causes the intended AppArmor or SELinux restrictions to not be applied to the container process, meaning the process runs with the same LSM labels as the Singularity runtime itself (typically unconfined for AppArmor, or container_runtime_t for SELinux). This results in low-to-moderate impacts on confidentiality, integrity, and availability, as the container process may perform operations that the LSM policy was intended to block. Critically, because Singularity runs non-root containers as the invoking non-root user, privilege escalation to root is not directly possible through this vulnerability alone; however, the bypass could be chained with other vulnerabilities (e.g., runc CVE-2025-52881) to achieve more severe outcomes (GitHub Advisory).

Exploitability

No public proof-of-concept exploit code or in-the-wild exploitation has been reported for CVE-2025-64750. The EPSS score is approximately 0.011% (0.000110), indicating very low probability of exploitation in the near term. The vulnerability is not listed in the CISA Known Exploited Vulnerabilities (KEV) catalog. Exploitation requires high attack complexity: the attacker must trick a user into running a malicious container image, control the content of a shared mount, and have knowledge of shared mount configurations on the target system (GitHub Advisory, Feedly).

Exploitation steps

  1. Craft a malicious container image: Build a Singularity container image that, upon execution, redirects the mount of /proc to the destination of a shared mount known to be configured on the target system (or one that will be specified by the user at runtime).
  2. Establish control of the shared mount: Either run a second malicious container that binds the same shared mount and controls its content, or leverage host-level permissions to control the content of the shared mount source.
  3. Induce the victim to run the container with LSM flags: Social-engineer or otherwise cause the target user to execute the malicious container image using singularity run --security apparmor:<profile> or --security selinux:<label>, which triggers the LSM label write to /proc/self/attr/exec.
  4. Redirect the LSM label write: The malicious container's /proc mount redirect causes Singularity to write the LSM label to a non-procfs file (e.g., a tmpfs file or a no-op procfs file), making the write ineffective.
  5. Container runs without LSM restrictions: The container process executes without the intended AppArmor or SELinux profile applied, allowing operations that the LSM policy was designed to prevent (GitHub Advisory, runc Advisory).

Indicators of compromise

  • Logs: Absence of expected AppArmor or SELinux denial log entries in /var/log/audit/audit.log or /var/log/syslog for container processes that should be confined; Singularity warning messages about SELinux/AppArmor being unavailable in container execution logs (pre-patch behavior).
  • Process: Container processes running with unconfined AppArmor label or container_runtime_t SELinux context instead of the user-specified profile, detectable via cat /proc/<pid>/attr/current or ps -Z.
  • File System: Unexpected shared mount configurations or bind mounts involving /proc within container namespaces; presence of unusual container images referencing shared mount paths known to the host system.
  • Network: No specific network indicators; exploitation is local and requires user interaction.

Mitigation and workarounds

Upgrade to SingularityCE 4.3.5 or SingularityPRO 4.1.11 / 4.3.5, which contain all three fixes: safe procfs API usage for AppArmor label writes (commit 5af3e79), updated containers/selinux dependency to v1.13.0 addressing CVE-2025-52881 (PR #3850), and conversion of silent LSM warnings to fatal errors (commit 2788296). There are no known configuration-based workarounds other than defining system-wide AppArmor/SELinux policy for Singularity itself, which applies to all containers regardless of --security flags and may impact functionality. Users should also avoid running container images from untrusted or unverified sources (GitHub Advisory).

Community reactions

The advisory was published by Sylabs maintainer dtrudg on December 2, 2025, and explicitly references the related runc vulnerability CVE-2025-52881, noting that the more severe container escape and denial-of-service attacks described in that advisory are not applicable to Singularity's non-root container model. Red Hat has acknowledged the CVE and SUSE issued a security update (SUSE-SU-2025:4395-1) addressing the issue in their distributions. No significant broader community or social media discussion has been identified beyond standard vulnerability tracking platforms (GitHub Advisory, Red Hat CVE).

Additional resources


SourceThis report was generated using AI

Related Linux Ubuntu vulnerabilities:

CVE ID

Severity

Score

Technologies

Component name

CISA KEV exploit

Has fix

Published date

CVE-2026-63343CRITICAL9.9
  • Linux Debian logoLinux Debian
  • incus
NoYesAug 21, 2026
CVE-2026-50538HIGH8.8
  • Linux Debian logoLinux Debian
  • veyon
NoYesAug 21, 2026
CVE-2026-53525HIGH7.4
  • Linux Debian logoLinux Debian
  • weechat
NoYesAug 21, 2026
CVE-2026-53524MEDIUM6.5
  • Linux Debian logoLinux Debian
  • weechat
NoYesAug 21, 2026
CVE-2026-44517MEDIUM6.3
  • Podman logoPodman
  • buildah-debuginfo
NoYesAug 21, 2026

Free Vulnerability Assessment

Benchmark your Cloud Security Posture

Evaluate your cloud security practices across 9 security domains to benchmark your risk level and identify gaps in your defenses.

Request assessment

Get a personalized demo

Ready to see Wiz in action?

"Best User Experience I have ever seen, provides full visibility to cloud workloads."
David EstlickCISO
"Wiz provides a single pane of glass to see what is going on in our cloud environments."
Adam FletcherChief Security Officer
"We know that if Wiz identifies something as critical, it actually is."
Greg PoniatowskiHead of Threat and Vulnerability Management