PyMuPDF versions up to 1.28.2 have a path traversal vulnerability in the font branch of the extract_objects() function located in src/__main__.py. The vulnerability arises because the output filename is constructed by concatenating a document-controlled BaseFont name directly onto a user-supplied output directory without sanitizing path separators or dot-dot sequences. An attacker can exploit this by supplying a crafted document (PDF, EPUB, XPS, or FB2) with a BaseFont name containing encoded path separators that decode to '../' sequences or absolute paths, enabling arbitrary file writes outside the intended directory. This flaw does not require authentication or elevated privileges to exploit. The vulnerability is tracked as CVE-2026-82035 and has a CVSS 3.1 score of 7.1 (high severity). The issue was fixed in commit b2c8f3a.

Successful exploitation allows an attacker to write arbitrary files to locations outside the intended output directory. This can lead to unauthorized file creation or modification, potentially impacting system integrity. The vulnerability does not affect confidentiality but has a high impact on integrity and a low impact on availability. No authentication or elevated privileges are required to exploit this vulnerability.

A fix for this vulnerability is available and was implemented in commit b2c8f3a. Users should update PyMuPDF to a version that includes this commit to remediate the vulnerability. Since this is not a cloud service, users must apply the patch themselves. Patch status is confirmed by the presence of the fixing commit. Until updated, avoid processing untrusted documents with affected versions.