The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill performs runtime compilation and library injection to bypass environment restrictions.
scripts/office/soffice.py contains a C source string (_SHIM_SOURCE) which is written to the temporary directory and compiled using gcc -shared -fPIC to create lo_socket_shim.so.
The resulting shared object is then injected into the LibreOffice process via the LD_PRELOAD environment variable to shim socket-related system calls (socket, listen, accept).
[COMMAND_EXECUTION]: The skill makes several calls to the system shell to manage Office processes and recalculate formulas.
scripts/recalc.py executes soffice with headless arguments and a custom macro to force formula evaluation.
scripts/office/soffice.py executes gcc to compile the socket shim.
scripts/office/validators/redlining.py executes git diff to perform document comparisons.
[INDIRECT_PROMPT_INJECTION]: The skill creates a surface for indirect injection by processing external data with high-privilege tools.
Ingestion points: Data enters the context via pd.read_excel() and load_workbook() calls described in SKILL.md.
Boundary markers: None identified; instructions do not specify delimiters for untrusted spreadsheet content.
Capability inventory: The skill can execute subprocesses (scripts/recalc.py, scripts/office/soffice.py), write to the filesystem (scripts/office/pack.py), and load arbitrary shared libraries.
Sanitization: No sanitization logic for ingested tabular data is present in the analyzed scripts.

xlsx