A vulnerability in pypdf versions prior to 6.21.3 can cause long runtimes when retrieving page labels from a crafted PDF file. This issue has been fixed in version 6.21.3. Users unable to upgrade immediately can apply the changes from the referenced pull request as a workaround.