docx

name	docx
description	Use this skill to read or create Word documents (.docx files). Supports text extraction and document creation with headings and lists.
metadata	{"yiyi":{"emoji":"📝","requires":{}}}

Word Document Processing

Drive everything through Python python-docx via run_python_script. If the package isn't present, call pip_install(["python-docx"]).

Reading DOCX Files

# read_docx.py
import sys
from docx import Document

doc = Document(sys.argv[1])
for p in doc.paragraphs:
    if p.text.strip():
        print(p.text)

For richer extraction (headings, lists, tables), inspect p.style.name and iterate doc.tables.

Creating DOCX Files

# create_docx.py
from docx import Document

doc = Document()
doc.add_heading("Report Title", 0)
doc.add_heading("Introduction", level=1)
doc.add_paragraph("This is the first paragraph.")
doc.add_paragraph("Item one", style="List Bullet")
doc.add_paragraph("Item two", style="List Bullet")
doc.add_heading("Conclusion", level=1)
doc.add_paragraph("Final remarks.")
doc.save("/path/to/output.docx")

Workflow Examples

Summarize a DOCX

1. Run the read script to extract text.
2. Analyze and summarize the content for the user.

Convert PDF to DOCX

1. Use the pdf skill to extract text via pypdf.
2. Use create_docx.py to write that text into a new Word document.

Create a report from data

1. Use the xlsx skill to read + aggregate the data.
2. Use create_docx.py to write a formatted report.

Advanced Operations

Accept All Tracked Changes (requires LibreOffice)

Remove tracked changes by accepting them all:

python3 scripts/accept_changes.py input.docx output.docx

Add Comments to DOCX

Add comments or replies to a DOCX document:

# First unpack the DOCX
python3 -c "from scripts.office.unpack import unpack; unpack('doc.docx', 'unpacked/')"

# Add a comment (id=0)
python3 scripts/comment.py unpacked/ 0 "Comment text"

# Add a reply to comment 0
python3 scripts/comment.py unpacked/ 1 "Reply text" --parent 0

# Repack
python3 -c "from scripts.office.pack import pack; pack('unpacked/', 'output.docx')"

Python python-docx (for tables, images, styles)

For complex documents with tables, images, page numbers, or tracked changes:

from docx import Document
doc = Document()
doc.add_heading("Report Title", 0)
doc.add_paragraph("Content paragraph...")
table = doc.add_table(rows=2, cols=3)
doc.save("output.docx")

Check: python3 -c "import docx; print('OK')"

Format Conversion (pandoc)

pandoc input.md -o output.docx      # Markdown → DOCX
pandoc input.docx -o output.md      # DOCX → Markdown
pandoc input.docx -o output.pdf     # DOCX → PDF

Quick Reference

Task	Approach
Read / extract text	Python python-docx (pip_install(["python-docx"]) if missing)
Create with headings/lists	Python python-docx
Accept tracked changes	scripts/accept_changes.py (requires LibreOffice)
Add comments	scripts/comment.py (requires unpack/repack)
Complex formatting	Python python-docx
Format conversion	pandoc CLI