The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script scripts/office/soffice.py (and its duplicate in the xlsx/ directory) includes an embedded C source string that is written to the temporary directory and compiled at runtime using gcc. The resulting shared library is then loaded into the LibreOffice process using the LD_PRELOAD environment variable to modify process behavior.
[COMMAND_EXECUTION]: The skill frequently uses subprocess.run to execute shell commands. This includes invoking gcc to compile the socket shim, calling soffice for spreadsheet recalculation in scripts/recalc.py, and using git diff for redlining validation in scripts/office/validators/redlining.py.
[PRIVILEGE_ESCALATION]: Through the use of LD_PRELOAD in scripts/office/soffice.py, the skill performs function hooking on critical system calls such as socket, listen, accept, and close. This allows the skill to intercept and redirect inter-process communication within the context of the LibreOffice application, effectively escalating its influence over external processes.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and process spreadsheet files provided by the user. Because these files are handled by scripts with extensive system access (including command execution and process manipulation), the skill presents a significant surface for indirect prompt injection if the data files contain malicious content designed to exploit the agent's capabilities.

xlsx