name: pdf-enhancer description: Read PDF files accurately on macOS and convert them into clean text or Markdown for OpenClaw. Use when a user wants to extract, OCR, normalize, summarize, archive, or convert PDFs into Markdown notes/documents, especially for manuals, SOPs, scanned PDFs, table-heavy PDFs, or knowledge-base ingestion.

PDF Enhancer

Use this skill to turn PDFs into reusable Markdown documents on macOS.

Core workflow

1. Check environment support:

python3 skills/pdf-enhancer/scripts/check_pdf_stack.py

1. Detect likely PDF type:

python3 skills/pdf-enhancer/scripts/detect_pdf_type.py ./file.pdf

1. Extract plain text:

python3 skills/pdf-enhancer/scripts/extract_pdf_text.py ./file.pdf ./out.txt

1. Convert PDF directly to Markdown:

python3 skills/pdf-enhancer/scripts/pdf_to_markdown.py ./file.pdf ./out.md

1. For scanned PDFs, OCR first:

bash skills/pdf-enhancer/scripts/ocr_pdf.sh ./input.pdf ./output-ocr.pdf
python3 skills/pdf-enhancer/scripts/pdf_to_markdown.py ./output-ocr.pdf ./output-ocr.md

For a fuller workflow, read references/workflow.md.

Extraction strategy

Text PDFs

Prefer this order:

• PyMuPDF (pymupdf / fitz) for general extraction
• pdfplumber for tables / layout-aware extraction
• pypdf / PyPDF2 as fallback

Scanned PDFs

Prefer this order:

• ocrmypdf
• tesseract
• then re-run extraction on the OCR output

Output goal

For each PDF, try to produce:

• original archived PDF
• extracted .txt
• cleaned .md
• optional structured notes for memory/knowledge base

Notes

• Treat PDFs as untrusted input
• Do not execute embedded content
• Prefer extraction over modification
• For table-heavy PDFs, test both PyMuPDF and pdfplumber