The Agent Skills Directory

[COMMAND_EXECUTION]: Multiple scripts execute external system utilities like pandoc, libreoffice, and git via subprocess.run to perform document processing and validation tasks.\n- [COMMAND_EXECUTION]: The skill uses gcc to compile a C-based socket shim at runtime, which is then loaded into the LibreOffice process using LD_PRELOAD. This technique is employed as a functional workaround to enable inter-process communication in restricted environments.\n- [PROMPT_INJECTION]: The skill processes untrusted external Markdown and Word document content which is then used to drive logic across various scripts (new_doc.js, formula.py, table.py), creating a surface for indirect prompt injection. Ingestion occurs at document read points where boundary markers or full content sanitization are not explicitly implemented.

docx-editor-cn