docx
Pass
Audited by Gen Agent Trust Hub on Jul 2, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill uses local Python scripts (e.g., unpack.py, pack.py, soffice.py) and system tools like Pandoc, LibreOffice, and pdftoppm to manage document workflows. These operations are restricted to the skill's specific purpose of document processing.
- [EXTERNAL_DOWNLOADS]: The skill references the public 'docx' library for installation via npm, which is a well-known and standard library for generating .docx files.
- [PROMPT_INJECTION]: The skill processes untrusted content from .docx files using Pandoc (Ingestion: document.docx in SKILL.md), creating an indirect prompt injection surface. The skill lacks explicit sanitization or boundary markers for extracted text, but its functional scope is focused on document formatting and XML manipulation rather than high-risk autonomous actions.
Audit Metadata