---

name: docx-tools description: Extract, inspect, and compare .docx files by examining their XML structure. Use when analyzing Word documents, comparing versions, or debugging document issues. allowed-tools: Bash, Read, Glob

DOCX Tools

Extract, inspect, and compare .docx files by examining their underlying XML structure. Since .docx files are ZIP archives containing XML, we use standard Unix tools.

Installation

Copy# macOS - tree is optional, other tools are pre-installed
brew install tree

# Ubuntu/Debian
sudo apt install unzip libxml2-utils tree

---

Quick Reference

---

Inspection Workflow

Step 1: List archive contents

Copyunzip -l document.docx

Step 2: View document structure (after extraction)

Copyunzip -o -d /tmp/docx-out document.docx
tree /tmp/docx-out

Step 3: View formatted XML content

Copyunzip -p document.docx word/document.xml | xmllint --format -

---

Comparison Workflow

Step 1: Quick structure check

Copyunzip -l document-v1.docx
unzip -l document-v2.docx

Step 2: Compare main document content

Copydiff -u <(unzip -p document-v1.docx word/document.xml | xmllint --format -) \
        <(unzip -p document-v2.docx word/document.xml | xmllint --format -)

Step 3: Compare styles if formatting changed

Copydiff -u <(unzip -p document-v1.docx word/styles.xml | xmllint --format -) \
        <(unzip -p document-v2.docx word/styles.xml | xmllint --format -)

---

Extraction Commands

Copy# Extract entire .docx to temp directory
unzip -o -d /tmp/docx-extracted document.docx

# Extract specific file only
unzip -o -d /tmp/docx-extracted document.docx word/document.xml

# Pipe extraction (no temp files) - view without extracting
unzip -p document.docx word/document.xml

# List archive contents with details
unzip -l document.docx

---

Comparison Commands

Copy# Unified diff (recommended)
diff -u <(unzip -p a.docx word/document.xml | xmllint --format -) \
        <(unzip -p b.docx word/document.xml | xmllint --format -)

# Side-by-side comparison
diff -y <(unzip -p a.docx word/document.xml | xmllint --format -) \
        <(unzip -p b.docx word/document.xml | xmllint --format -)

# Check if files differ (quiet mode)
diff -q <(unzip -p a.docx word/document.xml | xmllint --format -) \
        <(unzip -p b.docx word/document.xml | xmllint --format -)

# Compare file lists between documents
diff <(unzip -l a.docx | awk 'NR>3 {print $4}' | sort) \
     <(unzip -l b.docx | awk 'NR>3 {print $4}' | sort)

---

Inspection Commands

Copy# View metadata (author, dates)
unzip -p document.docx docProps/core.xml | xmllint --format -

# View app info (word count)
unzip -p document.docx docProps/app.xml | xmllint --format -

# Show structure (after extraction)
tree /tmp/docx-out

# Show only XML files
tree /tmp/docx-out -P '*.xml'

---

Key XML Files

---

Troubleshooting

Large diff output: Pipe to head -100 to limit lines.

Namespace noise: Focus on <w:t> tags for actual text content.

Binary comparison errors: Skip word/media/ directory - compare XML only.

Empty xmllint output: Check file path inside archive with unzip -l.