docx

Original：🇨🇳 Chinese

Translated

Creation, editing, and analysis of Word documents, supporting track changes, comments, format retention, and text extraction. Use this when you need to create .docx files, modify content, handle track changes/comments, or perform other document tasks.

1installs

Sourcekunhai-88/skills

Added on2026-02-17

NPX Install

npx skill4agent add kunhai-88/skills docx

SKILL.md Content (Chinese)

View Translation Comparison →

DOCX Creation, Editing, and Analysis

Overview

.docx is essentially a ZIP file containing resources such as XML. Select different workflows based on task types.

Workflow Selection

Read/Analysis Only → Text extraction or raw XML access
Create New Document → "Create New Word Document" workflow
Edit Existing Document:
- Your own document + simple modifications → Basic OOXML editing
- Documents from others / legal, academic, business, government documents → Redlining (Track Changes) Workflow (recommended or required)

Reading and Analysis

Text Extraction

Convert to markdown with pandoc, which can retain track changes:

bash

pandoc --track-changes=all path-to-file.docx -o output.md
# Options: --track-changes=accept/reject/all

Raw XML

Comments, complex formatting, structure, media, and metadata require unpacking to read XML.
Unpack:

python ooxml/scripts/unpack.py <office_file> <output_directory>

Key paths:

word/document.xml

,

word/comments.xml

,

word/media/

; Track changes use

<w:ins>

and

<w:del>

.

Create New Document

Use docx-js (JavaScript/TypeScript). First read docx-js.md thoroughly, then build with Document / Paragraph / TextRun, and export .docx using

Packer.toBuffer()

.

Edit Existing Document

Use Document Library (Python, operates on OOXML). Workflow:

Read ooxml.md thoroughly

Unpack:

unpack.py <office_file> <output_directory>

Write scripts to edit using the Document Library
Pack:
```
pack.py <input_directory> <office_file>
```

Redlining (Track Changes) Workflow

Markdown Representation:

pandoc --track-changes=all ... -o current.md

Identify and Batch Modifications: Group by chapter/type/difficulty, 3~10 changes per batch.
Unpack, read ooxml.md, use RSIDs as recommended.
Implement in Batches: Use
```
grep
```
to locate
```
word/document.xml
```
, use
```
get_node
```
etc. to implement changes,
```
doc.save()
```
.
Pack: Generate .docx with
```
pack.py
```
.
Verify: Convert to md again with
```
pandoc --track-changes=all
```
, use
```
grep
```
to check if modifications are complete and no extra changes exist.

Principle: Only mark actually changed text; reuse original

<w:r>

and RSIDs for unchanged parts.

Convert to Images