The Agent Skills Directory

[COMMAND_EXECUTION]: The recalc.py script uses subprocess.run to execute the soffice (LibreOffice) binary. This is used to recalculate Excel formulas in headless mode. The command is constructed as a list, which is a security best practice to prevent shell injection.
[PERSISTENCE]: The recalc.py script creates a LibreOffice macro file (Module1.xba) in the user's configuration directory (~/.config/libreoffice or ~/Library/Application Support/LibreOffice). This macro remains on the system across sessions to facilitate the RecalculateAndSave function. This is a legitimate functional requirement for the skill's purpose.
[DYNAMIC_EXECUTION]: The script generates executable Basic macro content at runtime and saves it to a file, which is then invoked by LibreOffice. This is used specifically to trigger the internal calculateAll() and store() methods of a spreadsheet document.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data from Excel, CSV, and TSV files.
Ingestion points: Data enters the agent context via pandas.read_excel and openpyxl.load_workbook as described in SKILL.md.
Boundary markers: The instructions do not specify strict delimiters for external data, though they emphasize the use of formulas over hardcoded values.
Capability inventory: The skill has the ability to write files (wb.save), execute system commands via recalc.py, and modify LibreOffice configurations.
Sanitization: There is no explicit sanitization of cell content before processing, which is standard for data analysis tools but constitutes a potential surface for indirect injection if data contains malicious instructions.

xlsx