About this skill

The Problem

When the main artifact is a Microsoft Word document or .docx file, the risk is usually not only incorrect text, but unstable OOXML structure. A .docx file is a ZIP package containing parts such as word/document.xml, styles.xml, numbering.xml, headers, footers, and relationship files. Visible text may be split across runs, revision tags, field boundaries, or bookmarks; treating it as plain text can silently damage styles, numbering, comments, cross-references, and table layout.

How It Works

This skill separates Word work into three paths: reading/extraction, generation, and preserving edits.

Reading prioritizes structure-preserving extraction for summaries, review, and content extraction.
Generation favors named styles over stacking direct formatting on body text.
Editing pays attention to abstractNum, num, sections, header/footer relationships, field caches, and revision metadata.

Key principles: list numbering is its own system; page layout is section-level; tracked changes, comments, and fields need minimal replacements; and round-trip compatibility across Word, LibreOffice, and Google Docs should be verified before delivery.

Boundaries

It fits business documents with styles, tables, headers, numbering, revisions, comments, and fields, especially where review context must be preserved. It is less necessary for plain-text export, simple conversion, or full rewrites. Legacy .doc and macro-bearing .docm files should be handled with conversion and risk assessment first.

Screenshots

Use Cases

Reviewing a contract, replace only disputed clauses while preserving Word revisions, comments, and bookmarks instead of rewriting whole paragraphs.
Maintaining a quote template, adjust headers, footers, A4/US Letter page size, and table widths to avoid layout drift in Google Docs.
Generating a report, use named styles and Word numbering instead of pasting Unicode bullets or spacing text manually.
Cleaning legacy files, inspect `word/document.xml`, `styles.xml`, `numbering.xml`, and field caches to catch hidden revisions.

Best For

Legal operations editing contract drafts, needing minimal replacements that preserve tracked changes and review metadata.
Product managers maintaining report templates, needing stable headers, styles, and tables across Word and LibreOffice.
Pre-sales consultants delivering client documents, checking fields, comments, cross-references, and pagination before handoff.
Technical support engineers debugging Word OOXML, locating `document.xml`, relationships, and numbering state issues.