Doc

DOCX

When to use

• Read or review DOCX content where layout matters.
• Create or edit DOCX files with professional formatting.
• Validate visual layout before delivery.

Workflow

If this skill is installed under DeepAgents, replace ~/.cline/skills in the examples below with ~/.deepagents/agent/skills.

1. Prefer visual review for layout, tables, diagrams, and pagination.
2. Use python-docx for edits and structured creation.
3. If the current Cline model cannot inspect images well enough, route rendered pages or the DOCX itself through onprem-document-vision.
4. Re-render and inspect after meaningful changes.
5. If visual review is not possible, extract text as a fallback and call out layout risk.

Rendering helper

Bundled helper:

python "$HOME/.cline/skills/doc/scripts/render_docx.py" /path/to/file.docx --output_dir /tmp/docx_pages

The helper renders DOCX through LibreOffice and rasterizes the resulting PDF with pypdfium2.

If the active model lacks vision, use the shared on-prem document vision helper directly:

python "$HOME/.cline/skills/onprem-document-vision/scripts/document_vision_review.py" `
  review "C:\path\to\file.docx" `
  --instructions "Extract visible text, explain diagrams, and call out layout issues." `
  --markdown

Dependencies

Python packages:

python -m pip install python-docx pypdfium2 pillow

Required system tools for rendering:

# macOS
brew install libreoffice

# Ubuntu or Debian
sudo apt-get install -y libreoffice

Quality expectations

• Deliver a polished document with consistent typography, spacing, margins, and hierarchy.
• Avoid clipped text, broken tables, unreadable characters, or placeholder styling.
• Re-render and inspect every page before final delivery when possible.
• If only text fallback was available, say that screenshots, diagrams, or layout-sensitive issues may remain.