The Agent Skills Directory

[COMMAND_EXECUTION]: The script resources/recalc.py uses subprocess.run to execute system commands including the LibreOffice soffice binary and system timeout utilities like timeout and gtimeout.- [DYNAMIC_EXECUTION]: The resources/recalc.py script dynamically generates a StarBasic macro file (Module1.xba) and then triggers its execution within the LibreOffice environment using the soffice command-line interface.- [PERSISTENCE]: The resources/recalc.py script modifies the user's LibreOffice application configuration directory (e.g., ~/.config/libreoffice/ or ~/Library/Application Support/LibreOffice/) to install a custom macro that persists across application sessions.- [INDIRECT_PROMPT_INJECTION]: The skill processes external document files such as XLSX and DOCX which may contain malicious instructions that could influence the agent's behavior.
Ingestion points: The resources/recalc.py script reads Excel files, and the skill documentation covers processing of PDF, DOCX, XLSX, and PPTX formats.
Boundary markers: The instructions lack specific delimiters or warnings for the agent to ignore instructions found within external files.
Capability inventory: The skill provides access to subprocess execution and file system write operations.
Sanitization: Content extracted from documents is not sanitized or validated before being used in the agent's context.- [EXTERNAL_DOWNLOADS]: The resources/docx-js.md file contains instructions for installing the docx package via npm. This refers to a well-known third-party library for document generation.

doc-writing