CVE-2026-94603

A flaw was found in Podman. When a container image with checkpoint annotations is executed using the podman run command, Podman treats the image as a restored checkpoint and ignores user-specified sandboxing options, such as dropped privileges. An attacker can exploit this issue by enticing a user to run a specially crafted image, leading to a container sandbox bypass and potential execution with elevated system privileges.

This vulnerability is rated as having an Important severity impact for Red Hat environments because running a malicious image can completely circumvent expected confinement boundaries and grant unexpected privileges to a container, compromising host integrity. Under default configurations, container execution relies on enforced sandboxing limits and security flags supplied by the operator. However, if an image embeds specific checkpoint metadata, the runtime automatically prioritizes the checkpoint configuration over explicit command-line restrictions, silently omitting requested security controls. Exploitation requires an operator or automated workflow to launch an untrusted image; environments that execute only vetted container workloads from trusted registries remain at reduced operational risk.

The CVSS score(s) listed for this vulnerability do not reflect the associated product's status, and are included for informational purposes to better understand the severity of this vulnerability.

Mitigation for this issue is either not available or the currently available options do not meet the Red Hat Product Security criteria comprising ease of use and deployment, applicability to widespread installation base, or stability.