The Agent Skills Directory

Dynamic Execution and Library Injection: The module scripts/office/soffice.py contains a C source code string that is written to a temporary file and compiled at runtime using gcc. The resulting shared object is then loaded into the soffice (LibreOffice) process via the LD_PRELOAD environment variable. This pattern is used to shim socket behavior for IPC in restricted environments, but runtime compilation and library injection are advanced techniques that warrant review.
Command Execution: The skill invokes several external command-line tools via subprocess.run(). This includes gcc for compilation, soffice for spreadsheet recalculation, and git diff for comparing document versions. While these tools are necessary for the skill's primary tasks, such direct command execution interacts closely with the host operating system.
Indirect Prompt Injection Surface: The skill processes user-supplied Office documents (XLSX, DOCX, PPTX) and parses their internal XML components using libraries like lxml. In scripts/office/validators/base.py, external XML files are ingested for validation. This ingestion point, combined with capabilities like system command execution, creates a potential surface for indirect prompt injection if malicious instructions are embedded within the processed document metadata or content.

xlsx