The Linux kernel Landlock subsystem contained a flaw that improperly handled files and directories opened from disconnected directories. When a sandboxed process accessed a bind mount that had been renamed or moved such that the target became invisible from the mount point, the kernel incorrectly accumulated access rights by walking the source filesystem hierarchy without accounting for the mount point. This caused inconsistent and potentially wider access results, allowing a sandboxed task with limited write and read permissions to acquire additional capabilities beyond those intended by its Landlock domain. The vulnerability represents an improper access control weakness that can lead to unintentional privilege escalation inside a sandboxed environment.

All installations of the Linux kernel that use the Landlock file‑system security module. The vulnerability affects any kernel configuration that enables Landlock; specific patched kernels are identified by the patch commit referenced in the advisory, but no version range is presently supplied in the official data.

The CVSS score of 5.5 indicates a medium severity. The EPSS score of less than 1% suggests that the probability of exploitation in the wild is currently very low. The vulnerability is not listed in the CISA KEV catalog. The likely attack vector is a local sandboxed task that has write access (FS_MAKE_REG, FS_REMOVE_FILE, FS_REFER) to the source of a bind mount and read access to the mount point; by performing a rename that crosses the mount boundary, the task could obtain a broader set of permissions than originally intended. Immediate patching of the kernel mitigates the risk, and in environments where patching is delayed, restricting the sandbox’s ability to create bind mounts or reducing its write privileges can help prevent the exploit. The mitigations are most effective when applied before any untrusted process is granted Landlock privileges that interact with bind mounts.