The Agent Skills Directory

[COMMAND_EXECUTION]: The helper script recalc.py executes the soffice (LibreOffice) binary and the gtimeout utility using subprocess.run. The commands are constructed using argument lists rather than shell strings, which follows safe implementation practices to prevent command injection.
[DYNAMIC_EXECUTION]: The skill performs runtime configuration by writing a StarBasic macro file (Module1.xba) to the user's LibreOffice configuration directory (~/.config/libreoffice/ or ~/Library/Application Support/LibreOffice/). This macro is generated from a hardcoded template to automate the recalculation and saving of workbooks.
[INDIRECT_PROMPT_INJECTION]: The skill processes external spreadsheet files, creating a vector for indirect instructions embedded in data.
Ingestion points: Untrusted spreadsheet data is read into the agent context via pd.read_excel() and load_workbook() as demonstrated in the skill instructions.
Boundary markers: The instructions do not implement explicit delimiters or instructions to the agent to disregard potential instructions within cell content.
Capability inventory: The skill grants the agent capabilities to write to the file system and execute shell commands through the recalc.py script.
Sanitization: No data validation or sanitization routines are specified for handling cell contents before they are used in further logic or outputs.

xlsx