word-track-changes（词轨变化）

Word Track Changes Skill

Enable native Microsoft Word Track Changes functionality using OpenXML. Handles real-world DOCX files where text is split across multiple <w:t> nodes.

When to Use

Use this skill when you need to:

• Edit a Word document while preserving a visible record of changes
• Show insertions as green underlined text
• Show deletions as red strikethrough text
• Allow users to accept/reject changes in Word's Review pane
• Replace text that may be split across multiple <w:r> / <w:t> elements

Quick Start

1. Enable Track Changes on a Document

python ~/.openclaw/skills/word-track-changes/scripts/enable_tracking.py input.docx output.docx

2. Replace Text with Revision Mark

from track_changes import TrackChangesProcessor

p = TrackChangesProcessor("document.docx")
p.set_author("Agent")
p.replace_text_with_revision(
    old_text="existing text to replace",
    new_text="new text with revision mark"
)
p.save("output.docx")
p.cleanup()

3. Insert a New Paragraph with Revision Mark

from track_changes import TrackChangesProcessor

p = TrackChangesProcessor("document.docx")
p.set_author("Agent")
p.insert_paragraph_after(
    search_text="text before new paragraph",
    new_text="New paragraph shown as insertion."
)
p.save("output.docx")
p.cleanup()

4. Batch Multiple Revisions

from track_changes import batch_revisions

revisions = [
    {"type": "replace", "old": "old text 1", "new": "new text 1"},
    {"type": "insert_after", "search": "anchor text", "new": "inserted paragraph"},
]

batch_revisions(
    "input.docx",
    revisions,
    author="Agent",
    output_path="output.docx"
)

Key Capabilities

Technical Details

OOXML Elements Used

• <w:ins> — Insertion mark (green, underlined)
• <w:del> — Deletion mark (red strikethrough)
• <w:delText> — Deleted text content

Architecture

DOCX (ZIP archive)
├── word/document.xml      ← Main content (modified with revision marks)
├── word/settings.xml      ← Track changes settings (updated if needed)
└── ...

The processor:

1. Unzips the DOCX into a temporary directory
2. Parses word/document.xml
3. Builds a paragraph-level text index to locate target text across split <w:t> nodes
4. Inserts <w:ins> / <w:del> at the exact position
5. Re-zips into a valid DOCX

Scripts Reference

All scripts are in scripts/ directory:

Python API

TrackChangesProcessor(docx_path)

Main class for processing a DOCX file.

Methods

• set_author(author) — Set the revision author name.
• replace_text_with_revision(old_text, new_text) — Replace text inside a paragraph with tracked changes. Raises ValueError if old_text is not found.
• insert_paragraph_after(search_text, new_text) — Insert a new paragraph after the paragraph containing search_text, marked as <w:ins>.
• enable_track_changes() — Update word/settings.xml to enable track revisions.
• save(output_path) — Write the modified DOCX.
• cleanup() — Remove temporary files.

Helper Functions

• enable_track_changes(input, output)
• insert_text_with_revision(input, search, new, author, output)
• mark_deletion(input, target, author, output)
• batch_revisions(input, revisions, author, output)

Example: Real-World Report Update

See example_usage.py for a complete working example.