The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes several internal Python scripts (such as scripts/office/unpack.py, scripts/office/pack.py, scripts/office/soffice.py, and scripts/comment.py) and system binaries (pandoc, pdftoppm) to perform document conversion, extraction, and editing tasks.
[EXTERNAL_DOWNLOADS]: The skill instructions specify the installation of the docx package from the npm registry using npm install -g docx.
[INDIRECT_PROMPT_INJECTION]: The skill operates on untrusted Word documents, creating a potential surface for indirect prompt injection attacks where content within the documents might influence the agent's behavior.
Ingestion points: Content is ingested from .docx files via pandoc text extraction and XML unpacking in SKILL.md.
Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded commands within the document text.
Capability inventory: The skill has capabilities for file writing (fs.writeFileSync), XML modification, and execution of various local processing scripts.
Sanitization: No sanitization or validation logic is specified for the text content extracted from external documents.

docx