PDF

PDF

When to use

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

• Read or review PDF content where layout matters.
• Create PDFs programmatically with reliable formatting.
• Validate final rendering before delivery.

Workflow

1. Use pdfplumber or pypdf first for text-heavy PDFs and quick checks.
2. If the PDF is scanned, image-heavy, table-heavy, or layout-sensitive, route it through onprem-document-vision instead of assuming the base LLM can inspect rendered pages.
3. Use reportlab to generate new PDFs.
4. After each meaningful update, re-render pages and verify alignment, spacing, and legibility.

Dependencies

Python packages:

python -m pip install reportlab pdfplumber pypdf pypdfium2 pillow

pypdfium2 handles page rendering directly, so no separate Poppler install is required for the shared PDF review path.

Rendering

If the active Cline model does not support vision well, use the shared on-prem vision helper:

python "$HOME/.cline/skills/onprem-document-vision/scripts/document_vision_review.py" `
  review "C:\path\to\file.pdf" `
  --max-pages 5 `
  --markdown

Quality expectations

• Keep typography, spacing, and hierarchy polished.
• Avoid clipped text, overlapping elements, broken tables, or unreadable glyphs.
• Do not deliver until the latest rendered pages are clean.
• If visual verification could not be completed through the on-prem vision helper, state that risk explicitly.