officecli
Pass
Audited by Gen Agent Trust Hub on Nov 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted Office documents (.docx, .xlsx, .pptx) which can contain embedded instructions designed to influence the agent's behavior during document analysis or modification tasks.\n
- Ingestion points: Untrusted data enters the agent context through document inspection and reading commands, specifically
officecli view,officecli get,officecli query, andofficecli dump(SKILL.md).\n - Boundary markers: The instructions do not define clear delimiters or provide the agent with warnings to ignore instructions found within the document content (SKILL.md).\n
- Capability inventory: The skill possesses significant capabilities to modify the file system and execute shell commands via the
officeclibinary, including operations likeadd,set,remove,move,swap, andbatch(SKILL.md).\n - Sanitization: There is no evidence of sanitization or validation of the document content before it is processed by the agent or outputted to its context.
Audit Metadata