gws-israeli-business-sheets
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill recommends installing the
@googleworkspace/clitool via NPM or from the official GitHub releases page. As this is a well-known tool from a reputable organization (Google Workspace) required for the skill's core functionality, this is documented neutrally as a safe practice.\n- [COMMAND_EXECUTION]: Thescripts/backup-sheets.pyutility usessubprocess.runto call thegwsCLI. The script correctly passes arguments as a list and does not enable shell execution, which is the recommended approach for secure command invocation.\n- [INDIRECT_PROMPT_INJECTION]: The skill acts as a data processing pipeline, reading financial records from Google Sheets and bank statements to generate reports. This creates an attack surface for indirect prompt injection if the processed data contains malicious natural language instructions.\n - Ingestion points: Spreadsheet data is retrieved via
gwsCLI calls (Step 5) and external files are loaded byscripts/vat-summary.py.\n - Boundary markers: The instructions do not define specific delimiters or instructions for the agent to ignore potentially malicious embedded content within the financial records.\n
- Capability inventory: The skill can read/write local files and manipulate Google Sheets via the authenticated CLI.\n
- Sanitization: While the Python scripts perform robust numeric parsing, they do not sanitize natural language text from the sheet cells for prompt injection threats.
Audit Metadata