CVE-2026-124424

AVideo translation tokens allowed subtitle writes outside the intended directory

WWBN · AVideo AI translation plugin

First documented disclosure · Review checkpoint

Summary

A valid AI job token could authorize writes to other subtitle locations accessible to the web service.

Severity and exploitation evidence

High · 7.1/10 · CVSS 3.1. Severity source.

Exploitation not established: The upstream report documents test writes. It does not identify malicious use or an affected customer.

Exploitation evidence source

Technical severity is not a measure of observed attacks or of risk to your deployment. Attempts, demonstrations, and successful compromise retain the qualifications in the source.

How the weakness works

The translation callback used untrusted language metadata to choose a subtitle destination and did not bind the operation to the correct owner.

Conditions required

An affected deployment has the bundled AI plugin enabled. The caller holds a valid AI job token and the web service can write the destination.

Potential impact

Subtitle files could be overwritten or created outside their intended directories. The reported writes were limited to the subtitle file extension.

Affected versions and configurations

The CNA lists through 29.7.0. The maintainer advisory identifies a checkout tested on October 9.

Fix and mitigation status

No confirmed patched release is named in the reviewed advisory or CNA record.

Practical next steps

Restrict access to the affected AI translation callback or disable the plugin pending a verified fix. Review subtitle writes and limit the service account’s filesystem permissions.

Operating lesson

Constrain a translation job’s output to that job’s owner and destination directory.

Evidence and limits

The upstream advisory describes verification on a test installation. Its proposed code changes do not identify a released fix.

Sources