office-document-review
Pass
Audited by Gen Agent Trust Hub on Aug 14, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and proofread text from external office documents, which may contain malicious instructions designed to influence the agent's behavior. Ingestion points: Target files in local or OneDrive directories including .docx, .doc, .odt, .rtf, .txt, and .pdf formats. Boundary markers: Absent; the instructions do not specify the use of delimiters or directives to ignore instructions within document content. Capability inventory: Local and cloud file system browsing, file reading, Python execution (zipfile, python-docx), and OCR tools. Sanitization: Absent; extracted text is processed directly without filtering or validation.
- [DYNAMIC_EXECUTION]: The skill provides logic for the agent to use Python modules like 'zipfile' and 'python-docx' to manually extract data from Word packages when standard readers fail, representing routine dynamic tool usage for specialized file handling.
- [COMMAND_EXECUTION]: The skill involves scanning and inventorying directory trees in the user's environment to resolve paths and locate target document folders.
Audit Metadata