This vulnerability involves improper limitation of a pathname to a restricted directory in the InstructLab component of RHEL AI 3. By exploiting a path traversal flaw in the chat session handler's 'logs_dir' parameter, a local attacker with low privileges can write files and create directories arbitrarily on the system. This can result in unauthorized data modification or disclosure. The CVSS 3.1 vector indicates local attack vector, low attack complexity, low privileges required, no user interaction, unchanged scope, and high impact on confidentiality and integrity but no impact on availability.

Successful exploitation allows a local attacker to write files and create directories outside the intended restricted directory, potentially leading to unauthorized data modification or disclosure. The confidentiality and integrity of the system can be compromised. There is no indication of impact on system availability. No known exploits in the wild have been reported at this time.

The vendor advisory does not explicitly state the availability of a patch or workaround. Patch status is not yet confirmed — check the Red Hat advisory at https://access.redhat.com/security/cve/CVE-2026-6855 for current remediation guidance. Until a fix is confirmed, restrict local access to trusted users only and monitor for suspicious activity related to file and directory creation in the affected component.