docx-official
Pass
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements multiple layers of security when handling potentially untrusted Office documents.
- [XML Security]: All XML parsing in the primary workflow is handled via 'defusedxml', providing comprehensive protection against XML External Entity (XXE) and Billion Laughs expansion attacks. Even where 'lxml' or 'xml.etree' are used for validation, the content has been pre-sanitized during the unpacking stage.
- [Archive Protection]: The 'unpack.py' script includes deterministic checks to prevent symlink-based traversal and ZIP member path escaping (ZipSlip), ensuring that documents are extracted safely into designated temporary directories.
- [Sanitized Commands]: Command execution via 'subprocess.run' is strictly controlled. It is used exclusively for headless LibreOffice conversion to validate document integrity and Git for word-level content comparison. These executions are local-only and do not include shell-injection vectors or network access.
- [Metadata Sanitization]: The 'document.py' library uses 'html.escape' for author metadata when updating 'people.xml', preventing structural XML injection vulnerabilities.
Audit Metadata