The Agent Skills Directory

[SAFE]: The skill serves as a documentation and utility wrapper for the official Microsoft markitdown tool. All Python code snippets provided are standard examples of document conversion and file I/O.
[EXTERNAL_DOWNLOADS]: The skill mentions installing the markitdown package via pip install markitdown. This is a well-known, legitimate library from a trusted organization (Microsoft).
[COMMAND_EXECUTION]: Provides command-line examples for using the library (markitdown document.docx > output.md). These are standard usage patterns for the documented tool.
[DYNAMIC_EXECUTION]: No dynamic code execution (eval/exec) or unsafe deserialization patterns were detected.
[INDIRECT_PROMPT_INJECTION]: While the skill processes external documents (PDF, Docx, etc.), which is an attack surface, the provided logic does not perform unsafe interpolation into LLM prompts without structure, and it does not grant the agent high-privilege capabilities that would be exploitable via such an injection.

office-to-md