docx Skill for AI Agents

# DOCX Processing

Comprehensive toolkit for Word document manipulation.

## Core Capabilities

- Document creation from scratch
- Editing existing documents
- Tracked changes support
- Comments and annotations
- Formatting preservation
- Text extraction and analysis

## Primary Workflows

### Reading/Analysis

Use text extraction for content analysis. Access raw XML for:
- Comments
- Complex formatting
- Document metadata

### Creating Documents

Use docx-js library with JavaScript/TypeScript to build new Word documents from scratch.

### Editing Existing Documents

Python-based Document library handles OOXML manipulation with:
- High-level methods for common operations
- Direct DOM access for complex changes

### Document Review (Redlining)

For tracked changes:
1. Plan changes in markdown first
2. Implement systematic OOXML edits
3. Preserve revision history

## Key Implementation Principles

### Minimal, Precise Edits

Only mark text that actually changes. Don't replace entire sentences when modifying a word.

### Preserve Unchanged Content

Reuse original XML elements with their RSID values for unchanged portions.

### Batching Strategy

Group 3-10 related changes per batch:
- Enables manageable debugging
- Maintains efficiency
- Reduces error surface

## Technical Operations

```
# Unpack document
python ooxml/scripts/unpack.py document.docx

# Modify XML files...

# Repack document
python ooxml/scripts/pack.py document.docx
```

## Visual Analysis

Convert documents to PDF then JPEG for visual inspection when needed.

## Best Practices

- Always work on copies, not originals
- Validate document structure after edits
- Test tracked changes display in Word
- Preserve styles and formatting