docx
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS] (INFO): References the
docxNode.js package and system utilities likepandoc. These dependencies are standard for the skill's stated purpose and are managed within a trusted repository environment (anthropics/skills). - [COMMAND_EXECUTION] (SAFE): Employs common CLI tools (
unzip,zip,soffice,pdftoppm) to manipulate document structures and handle format conversions. These actions are necessary for the skill's core functionality. - [PROMPT_INJECTION] (LOW): Identifies an indirect prompt injection surface (Category 8) due to the ingestion of untrusted document content. Evidence: 1. Ingestion points:
pandoctext extraction andunzip(parsingword/document.xml). 2. Boundary markers: Absent. 3. Capability inventory: File system writes (fs.writeFileSync) and shell command execution (zip,soffice). 4. Sanitization: Absent. The skill does not explicitly sanitize or delimit content extracted from the DOCX files before the agent processes it.
Audit Metadata