office-mcp
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it is designed to ingest and process untrusted data from various document formats (PDF, Word, Excel, PowerPoint). While this is inherent to document processing, it allows external content to enter the agent context.
- Ingestion points: Data ingestion occurs through extraction tools such as
extract_text_from_pdf,read_xlsx,extract_text_from_docx, andextract_from_pptxdescribed in SKILL.md. - Boundary markers: The skill instructions do not explicitly provide delimiters or specific instructions for the agent to treat extracted content as untrusted.
- Capability inventory: The skill has broad file system capabilities, including creating, modifying, and converting files across multiple formats (Word, Excel, PPTX, PDF, CSV, JSON, Markdown).
- Sanitization: No explicit sanitization or validation of the extracted document content is described in the skill metadata.
Audit Metadata