The vulnerability lies in incorrect handling of JPEG 2000 images within PDFs. Improper processing of low-bit-depth samples and mismatched component resolutions can cause Ghostscript to write 1,024 bytes into a 256-byte buffer, corrupting memory and enabling remote code execution.
Wordfence researchers, using their Argus AI-powered system, developed two proof-of-concept exploits that achieve unauthenticated remote code execution in a WordPress deployment. However, success depends on the target environment. They tested ten Linux distributions—including Debian, Ubuntu, Fedora, Alpine, and AlmaLinux—alongside an official WordPress Docker image snapshot, and found all contained the vulnerable code.
In lab tests, the exploits failed to achieve remote code execution in 10,240 attempts against standard Linux packages but succeeded on a purpose-built test server. Modifying the PDF's fonts, language settings, font configuration, and one core system library improved reliability.
For WordPress environments specifically, four conditions must be met: an attacker can upload a PDF (anonymously or via a low-privileged account); the site processes the PDF to generate a preview or thumbnail; the processing chain reaches Ghostscript with necessary software and policy settings; and Ghostscript contains the unpatched vulnerable decoding code. Stock WordPress does not allow anonymous Media Library uploads. More broadly, any application that renders attacker-supplied PDFs through affected Ghostscript may be exposed.
Artifex, the maintainer of Ghostscript, released a fix in version 10.08.0 on September 8. Fixes were also backported to older versions. WordPress administrators should ensure through their hosting provider that Ghostscript has been upgraded to version 10.08.0 or later. Docker users should rebuild and redeploy with an updated base image. If upgrading is not possible, blocking PDF uploads or disabling PDF rendering in ImageMagick is recommended.
Wordfence has not observed any exploitation in the wild so far.