Export

/export — Document Export Skill

Purpose

Convert thesis chapters and reading notes from Markdown to Word (.docx) format and package them into a ZIP archive for submission to supervisors or examiners. This skill is user-invoked only and will not be triggered automatically.

Trigger Words

This skill activates on: /export. It is never invoked automatically by the model.

Parameters

• $ARGUMENTS[0] -- scope: chapters | notes | all (default: all)
• $ARGUMENTS[1] -- language filter: en-only | all (default: all)

Examples:

• /export chapters en-only -- export only chapter files, skip files with significant CJK content
• /export notes all -- export only reading notes, include all languages
• /export -- export everything, all languages

Workflow

1. Check dependencies. Verify that pypandoc is installed AND a working pandoc binary is on PATH (the script smoke-probes via pypandoc.get_pandoc_version()). If pypandoc/pandoc are unavailable, fall back to python-docx + markdown. Report which conversion method is being used.

   Run from the project root.

2. Run the conversion script.

   python {skill_dir}/scripts/convert_to_docx.py \
     --base-dir {project_root} \
     --output-dir {project_root}/final_output \
     --scope {scope} \
     --lang-filter {lang_filter}
   

3. Report results.

   ## Export Complete
   
   | Item | Value |
   |------|-------|
   | Files converted | {N} |
   | Files skipped (language filter) | {N} |
   | Output directory | {path} |
   | ZIP archive | {path} |
   | Total size | {size} |
   
   Conversion method: {pandoc | python-docx fallback}
   

Constraints

1. User-invoked only. This skill has disable-model-invocation: true and must not run unless the user explicitly calls /export.
2. No hardcoded paths. All paths are derived from arguments or project root.
3. No emoji in output.
4. Preserve formatting as much as possible during conversion -- headings, tables, block quotes, inline code.