DOCX Brief Maker
Overview
Extract .docx content with WordprocessingML structure, then produce a rule-based brief that Codex can refine into higher-quality summaries, action plans, or edit recommendations.
Workflow
- Confirm the input is a local
.docxfile supplied or approved by the user. - Run
scripts/extract_docx_brief.pywith--format markdownor--format json. - Use the structured extraction as source material for any deeper Codex summary or rewrite.
- Keep page-layout claims modest; this script reads document XML and does not perform visual render QA.
- When the user needs polished document editing or final DOCX generation, use a document editing workflow after the brief is created.
Quick Start
python3 skills/docx-brief-maker/scripts/extract_docx_brief.py input.docx --format markdown --output brief.md
python3 skills/docx-brief-maker/scripts/extract_docx_brief.py input.docx --format json --output brief.json
Script Behavior
scripts/extract_docx_brief.py:
- reads
.docxwith Python standard-libraryzipfileand XML parsing; - extracts paragraphs, headings, list-like paragraphs, and table cell text;
- returns
summary,action_items,suggested_edits,risks_or_questions,source_metadata, and raw extracted structure; - produces deterministic rule-based brief fields;
- leaves complex judgment, nuanced editing, and polished writing to Codex using the extracted structure.
Reference
Read references/brief-guidelines.md for brief quality rules, action-item extraction cues, risk detection, and limitations.