microsoft-word

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill employs standard and well-maintained Python libraries (python-docx, docxtpl) for document manipulation, following established development practices.\n- [SAFE]: File system operations are confined to the expected tasks of reading template files, writing generated documents, and extracting embedded media from Word files.\n- [SAFE]: Network operations are restricted to interactions with the official Microsoft Graph API for document management and conversion, with no evidence of unauthorized data exfiltration to third-party domains.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources such as CSV files and existing Word documents to generate output, which is a standard document processing function.\n
  • Ingestion points: CSV data in Step 2 (clients.csv), existing Word files in Step 3 (report.docx), and user-supplied text in the provided examples.\n
  • Boundary markers: No specific delimiters or instructions to ignore embedded commands are present in the template rendering logic.\n
  • Capability inventory: Package installation (pip install), file system write access (doc.save), and network communication (Microsoft Graph API).\n
  • Sanitization: Input data is interpolated directly into templates without additional sanitization layers, which is common for document generation utilities.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 09:38 PM
Security Audit — agent-trust-hub — microsoft-word