The vulnerability arises from an overly permissive CSP implementation in Yelp's yelp-xsl stylesheet, which enables a malicious Flatpak application to bypass sandbox restrictions by supplying crafted help content through the OpenURI portal. This crafted content includes untrusted CSS stylesheets embedded in SVG documents that cause Yelp to evaluate local XML inclusions and disclose arbitrary user-readable files on the host system. Exploitation does not require privileges on Yelp or additional user interaction beyond running the Flatpak application. This cross-boundary information disclosure vulnerability was introduced as a regression related to a previous CVE (CVE-2025-3155) and is remediated solely by updating the yelp package. The vulnerability affects Red Hat Enterprise Linux 8 and related builds that include the vulnerable yelp package. Red Hat has released a security update (RHSA-2026:47177) that fully addresses this issue.

The vulnerability allows a sandboxed Flatpak application to bypass intended sandbox isolation and cause the Yelp application to disclose arbitrary user-readable files from the host system. This results in unauthorized disclosure of sensitive information (confidentiality impact is high). There is no impact on integrity or availability. Exploitation requires local access to run a malicious Flatpak application but no additional privileges or user interaction. This is a cross-boundary information disclosure vulnerability leveraging the host's Yelp application.

A security update for the yelp package is available and should be applied to fully remediate this vulnerability. No other mitigations meeting Red Hat's standards for usability and stability are currently available. Systems are fully protected once the corresponding yelp security update is installed. Users should monitor Red Hat advisories and apply the update as soon as possible.