FreeBSD EUVD-2026-101866

Exploitation requires: (1) attacker-controlled code already running inside a FreeBSD jail; (2) that jailed process must have received a directory file descriptor originating from another jail; and (3) the referenced directory must reside on a filesystem that is also reachable from the exploiting jail's root. Limiting factors that reduce real-world risk: it is not remotely exploitable (local in-jail execution is a prerequisite), the cross-jail descriptor-passing arrangement is a non-default/specialized deployment pattern, and the shared-filesystem reachability requirement further narrows applicability. Only patched releases (15.0-RELEASE p14, 14.5-RELEASE p1, 14.4-RELEASE p10, 15.1-RELEASE p4 and later) enforce FD_RESOLVE_BENEATH.

The input assigns CVSS 8.8 with AV:N (network) and AC:L, but the description describes a purely local jail-escape primitive: a process already executing inside a FreeBSD jail must receive a directory file descriptor from another jail and then abuse renameat(2)/fchdir(2). The attack vector is therefore local, not network - CVSS AV:N appears incorrect and should be verified against the FreeBSD advisory; this materially inflates the published score. Supporting signals point to modest real-world risk: EPSS is low at 0.18% (7th percentile), there is no CISA KEV listing (not actively exploited), and no public exploit identified at time of analysis. The CWE-273 (improper privilege/confinement handling) and the high C/I/A impacts are credible because a successful escape grants access to the host filesystem beyond the jail root. Net assessment: this is a genuine, impactful sandbox-confinement bypass for multi-tenant FreeBSD jail deployments and should be patched, but it is NOT an opportunistic, widely-exploitable network bug - the AV:N label overstates urgency, and the specific precondition (a directory FD passed in from another jail on a shared filesystem) makes it a targeted, configuration-dependent issue rather than a mass-exploitation threat.

In a multi-tenant FreeBSD host, an attacker controlling a process in one jail obtains a directory file descriptor that was passed in from another jail (e.g., via a shared descriptor-passing mechanism) on a filesystem also mounted under the jail root. Because renameat(2) does not enforce FD_RESOLVE_BENEATH, the attacker renames/relocates that directory outside the jail's confined subtree and then uses fchdir(2) to step into it, escaping the jail's filesystem-root restriction and reaching host files. No public exploit code is identified at time of analysis, and exploitation requires local code execution already inside a jail.