xlsx

Pass

Audited by Gen Agent Trust Hub on Feb 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The recalc.py script uses the subprocess module to invoke the soffice command from LibreOffice for headless formula recalculation.
  • Evidence: The recalc function in the provided script constructs a command list starting with 'soffice' and executes it using subprocess.run().
  • [COMMAND_EXECUTION]: The skill dynamically creates and manages a LibreOffice macro to automate workbook operations such as calculation and saving.
  • Evidence: The setup_libreoffice_macro function in recalc.py writes XML macro content to a Module1.xba file within the user's LibreOffice configuration directory.
  • [PROMPT_INJECTION]: The skill presents a potential surface for indirect prompt injection by reading external spreadsheet data while maintaining command execution capabilities.
  • Ingestion points: External data is ingested through pd.read_excel() and openpyxl.load_workbook() as demonstrated in the SKILL.md examples.
  • Boundary markers: The provided prompts and scripts do not include specific delimiters or instructions to the agent to disregard embedded commands in the processed data.
  • Capability inventory: The skill has the capability to execute the soffice binary and write to the local file system.
  • Sanitization: No specific content-based sanitization or filtering is performed on spreadsheet values before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 27, 2026, 08:37 AM