The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script scripts/office/soffice.py contains functionality to dynamically generate C source code at runtime and compile it into a shared object file using the system's gcc compiler.
This library is then injected into the LibreOffice (soffice) process using the LD_PRELOAD environment variable to shim network socket calls.
[COMMAND_EXECUTION]: The skill frequently invokes external system binaries to perform its tasks:
scripts/recalc.py executes soffice (LibreOffice) to perform headless formula recalculations.
scripts/office/soffice.py executes gcc to compile the compatibility shim.
scripts/office/validators/redlining.py executes git to perform word-level diffing between document versions.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a large attack surface for indirect injection as it processes untrusted spreadsheet data and has significant local execution capabilities.
Ingestion points: Reads and processes external .xlsx, .csv, and .tsv files provided by the user.
Boundary markers: None explicitly defined for isolating cell text from agent instructions.
Capability inventory: Subprocess execution of soffice, gcc, and git across multiple helper scripts.
Sanitization: None performed on the spreadsheet content before processing.

xlsx