The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts build_le_docx.py and make_le_template.py execute the pandoc command via subprocess.run. This is a routine operation for document conversion and template extraction within a build pipeline. The commands are well-defined and do not execute arbitrary shell strings.
[INDIRECT_PROMPT_INJECTION]: The skill reads user-provided markdown papers as input. While this exposes a surface for indirect prompt injection, the risk is minimal as the processing is restricted to formatting, citation resolution, and XML modification for styling. There is no evidence of the agent executing embedded instructions from the input data during the build process.

law-econ-docx