office-productivity
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The workflows for document creation, conversion, and automation create a surface for indirect prompt injection attacks.
- Ingestion points: The skill instructs the agent to process various external file formats (DOCX, XLSX, ODT, ODS) and data from Google Workspace.
- Boundary markers: No explicit delimiters or instructions are provided to help the agent distinguish between untrusted document content and legitimate instructions.
- Capability inventory: The workflow involves file system operations and API interactions for document automation and distribution.
- Sanitization: There is no requirement for the validation or sanitization of content within the processed documents.
Audit Metadata