audit-xls
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided spreadsheet data which could contain embedded instructions intended to manipulate the agent's logic.
- Ingestion points: Excel file content and range selections processed via pandas and openpyxl (SKILL.md).
- Boundary markers: None present to distinguish between data and instructions.
- Capability inventory: Execution of Python scripts for data analysis and shell commands for package management.
- Sanitization: No evidence of sanitizing or escaping cell content before analysis.
- [EXTERNAL_DOWNLOADS]: The skill uses pip to download and install 'pandas' and 'openpyxl' if they are missing from the environment.
- [COMMAND_EXECUTION]: Shell commands are executed to check for the presence of dependencies and perform installations during the preflight stage.
Audit Metadata