The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script scripts/office/soffice.py performs runtime compilation of C source code into a shared object using gcc. This library is then injected into the LibreOffice process using the LD_PRELOAD environment variable. This technique allows for arbitrary code execution and the modification of system-level function calls.
[PERSISTENCE_MECHANISMS]: The script scripts/recalc.py writes a StarBasic macro to the LibreOffice user configuration directory (~/.config/libreoffice/4/user/basic/Standard/). This macro remains in the user profile across sessions and is executed by LibreOffice to perform recalculations.
[COMMAND_EXECUTION]: The skill uses subprocess.run across multiple scripts to execute external system tools including LibreOffice (soffice), the GCC compiler (gcc), and Git (git).
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted Excel and Office files, creating an indirect prompt injection surface.
Ingestion points: Office XML components are parsed in scripts/office/unpack.py, and Excel files are loaded in scripts/recalc.py.
Boundary markers: Content is processed without specific delimiters or instructions for the agent to ignore embedded instructions.
Capability inventory: The skill has the ability to execute shell commands and compile/load system libraries.
Sanitization: Standard XML security measures are applied using the defusedxml library.

xlsx