A trust-boundary flaw in CRI-O's sandbox state persistence allows attacker-influenced pod metadata to overwrite CRI-O's own reserved sandbox bookkeeping; once reloaded as trusted after a restart, a later container recreate in that sandbox can expose a host-side runtime-management resource inside the container, enabling container escape.
There is no complete workaround; until a fix is available, administrators should restrict who can create/update pods and set arbitrary annotations via RBAC/admission policy, minimize unnecessary CRI-O restarts or node reboots and monitor for anomalous container mounts (e.g. under /dev/shm) following runtime restarts, then apply the vendor patch once released.