The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script scripts/office/soffice.py contains hardcoded C source code that is written to a temporary location and compiled using gcc at runtime. The resulting shared object is injected into the LibreOffice process via the LD_PRELOAD environment variable to override standard socket behavior.
[DYNAMIC_EXECUTION]: The script scripts/recalc.py automatically writes a StarBasic macro (Module1.xba) to the user's LibreOffice configuration directory (e.g., ~/.config/libreoffice) to enable automated formula recalculation.
[COMMAND_EXECUTION]: Multiple Python scripts utilize the subprocess module to execute shell commands, including gcc for code compilation, git for document comparisons, and the soffice binary for processing Excel files.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external spreadsheet files using pandas and openpyxl, which creates a potential vector for malicious instructions to influence the agent. Ingestion points: read_excel and load_workbook calls in primary workflows. Boundary markers: No explicit delimiters or instructions to ignore embedded commands are used when interpolating file data. Capability inventory: The skill possesses file write access and the ability to execute shell commands. Sanitization: Ingested cell data is not filtered or sanitized before processing.

xlsx