---

name: docx description: Create, edit, and analyze Word documents (.docx). Activates for "Word document", "docx", "write document", "create document", "edit document", "tracked changes", "redline". allowed-tools: [Read, Write, Bash, Task]

Word Document (DOCX) Skill

When This Skill Activates

• "Create a Word document", "write a docx"
• "Edit this document", "modify the Word file"
• "Add tracked changes", "redline this document"
• "Extract text from docx", "read this Word file"
• "Convert document to markdown"

Core Workflows

1. Reading/Analyzing Documents

Extract text to markdown:

Copy# Using pandoc (preferred)
pandoc input.docx -o output.md

# Using markitdown
python -m markitdown file.docx

Access raw XML for advanced features:

Copy# Unzip to access document.xml
unzip -o document.docx -d docx_extracted
cat docx_extracted/word/document.xml | xmllint --format -

2. Creating New Documents

Using python-docx:

Copyfrom docx import Document
from docx.shared import Inches, Pt
from docx.enum.text import WD_ALIGN_PARAGRAPH

doc = Document()

# Add title
title = doc.add_heading('Document Title', level=0)
title.alignment = WD_ALIGN_PARAGRAPH.CENTER

# Add paragraph with formatting
para = doc.add_paragraph()
run = para.add_run('Bold text')
run.bold = True
para.add_run(' and normal text.')

# Add table
table = doc.add_table(rows=3, cols=3)
table.style = 'Table Grid'
for i, row in enumerate(table.rows):
    for j, cell in enumerate(row.cells):
        cell.text = f'Row {i+1}, Col {j+1}'

# Add image
doc.add_picture('image.png', width=Inches(4))

doc.save('output.docx')

3. Editing Existing Documents

Surgical text replacement:

Copyfrom docx import Document

doc = Document('input.docx')

for para in doc.paragraphs:
    if 'OLD_TEXT' in para.text:
        for run in para.runs:
            if 'OLD_TEXT' in run.text:
                run.text = run.text.replace('OLD_TEXT', 'NEW_TEXT')

doc.save('output.docx')

Preserve formatting during edits:

Copy# CRITICAL: Only modify run.text, never recreate runs
# This preserves font, size, color, bold, italic, etc.
for run in para.runs:
    run.text = modified_text  # Keeps all formatting

4. Tracked Changes (Redlining)

Workflow for document review:

1. Convert to markdown for analysis
2. Plan changes systematically
3. Apply in batches (3-10 related changes)
4. Verify all changes applied

Batch strategy example:

CopyCHANGES = [
    {"section": "Introduction", "original": "...", "revised": "..."},
    {"section": "Methods", "original": "...", "revised": "..."},
]

for change in CHANGES:
    apply_change(doc, change)
    verify_change(doc, change)

Critical Requirements

Dependencies

Copypip install python-docx markitdown
brew install pandoc  # or apt-get install pandoc

Common Patterns

Add Header/Footer

Copysection = doc.sections[0]
header = section.header
header.paragraphs[0].text = "Header Text"

Page Break

Copyfrom docx.enum.text import WD_BREAK
doc.paragraphs[-1].runs[-1].add_break(WD_BREAK.PAGE)

Set Margins

Copyfrom docx.shared import Inches
section = doc.sections[0]
section.left_margin = Inches(1)
section.right_margin = Inches(1)

Output Format

Copy## DOCX Operation: [Create/Edit/Analyze]

### Input
[Source document or requirements]

### Actions Taken
1. [Action 1]
2. [Action 2]

### Output
[Result file path or extracted content]

### Verification
[Confirmation of success]