docx
Comprehensive document creation, editing, and analysis with support for tracked changes, comments, formatting preservation, and text extraction. When Claude needs to work with professional documents (.docx files) for: (1) Creating new documents, (2) Modifying or editing content, (3) Working with tracked changes, (4) Adding comments, or any other document tasks
87
Quality
85%
Does it follow best practices?
Impact
91%
1.62xAverage score across 3 eval scenarios
Evaluation results
100%
32%Employee Onboarding Guide
docx-js document creation rules
No unicode bullets
100%
100%
LevelFormat.BULLET constant
0%
100%
ShadingType.CLEAR for table
100%
100%
Table column widths
25%
100%
PageBreak inside Paragraph
100%
100%
Default font set
0%
100%
Heading style IDs
100%
100%
outlineLevel in heading styles
0%
100%
TOC uses HeadingLevel only
100%
100%
No \n in TextRun
100%
100%
Borders on cells not table
100%
100%
Concise code
100%
100%
Without context: $0.5463 · 2m 14s · 27 turns · 34 in / 8,652 out tokens
With context: $0.8046 · 2m 57s · 26 turns · 31 in / 10,966 out tokens
90%
36%Vendor Contract Review
Redlining workflow with tracked changes
Pandoc conversion step
100%
100%
Unpack.py used
0%
100%
PYTHONPATH set to skill root
0%
75%
Document library imported
0%
100%
Minimal precise edits
100%
100%
Grep before scripting
0%
0%
Pack.py used
0%
100%
All six changes applied
100%
100%
Pandoc final verification
100%
100%
Tracked change author
100%
100%
doc.save() called
0%
100%
No markdown line numbers for location
100%
100%
Without context: $0.4040 · 1m 50s · 21 turns · 21 in / 6,592 out tokens
With context: $1.6053 · 3m 21s · 45 turns · 10,820 in / 11,308 out tokens
85%
39%Document Audit Report
Document reading and XML analysis
Pandoc for text extraction
0%
80%
Pandoc --track-changes flag
0%
100%
Unpack.py for XML access
0%
0%
word/document.xml accessed
30%
100%
w:ins and w:del tag identification
100%
100%
word/comments.xml accessed
20%
100%
Correct headings in report
100%
100%
Tracked changes counts
100%
100%
Text preview present
100%
100%
Without context: $0.4588 · 1m 39s · 28 turns · 34 in / 6,047 out tokens
With context: $1.0911 · 2m 59s · 41 turns · 4,997 in / 9,549 out tokens
- Evaluated
- Agent
- Claude Code
- Model
- Claude Sonnet 4.6
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.