xlsx
Warn
Audited by Gen Agent Trust Hub on Mar 3, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script
recalc.pyprogrammatically interacts with the system to perform formula recalculation. - The script uses
subprocess.runto execute thesoffice(LibreOffice) binary in headless mode. - It utilizes the
vnd.sun.star.scriptprotocol to trigger the execution of a locally stored macro. - It dynamically writes a StarBasic macro (
Module1.xba) to the user's LibreOffice application support directory (~/Library/Application Support/LibreOffice/...on macOS or~/.config/libreoffice/...on Linux). - [PROMPT_INJECTION]: The skill exhibits a significant surface area for indirect prompt injection.
- Ingestion points: The skill reads external data from
.xlsx,.xlsm,.csv, and.tsvfiles usingpandas.read_excelandopenpyxl.load_workbookas described inSKILL.md. - Boundary markers: There are no explicit instructions or delimiters defined to prevent the agent from interpreting instructions embedded within the spreadsheet data.
- Capability inventory: The skill possesses powerful capabilities including file writing (
wb.save,macro_file.write) and system command execution (subprocess.runinrecalc.py). - Sanitization: There is no evidence of data validation or sanitization for content read from external spreadsheets before it is processed by the agent or used in calculations.
Audit Metadata