Claude Official Skill: PDF — Read, Create & Edit PDFs

What it is

This is an official Claude Code skill from Anthropic's skills repository for working with PDF files. It handles reading content, extracting text and tables, creating new PDFs, merging multiple documents, splitting pages, and converting between formats.

The skill is designed for developers and knowledge workers who use Claude Code to process PDF documents without switching to external tools or writing custom extraction code.

How it saves time or tokens

The skill activates automatically when you ask Claude Code to work with PDF files. Instead of researching which Python library to use (PyPDF2, pdfplumber, reportlab, pypdf), the skill provides Claude Code with the right approach for each task. The estimated token cost is around 500 tokens per invocation.

How to use

1. Install via Claude Code:

claude skill install anthropics/skills/pdf

1. Or manually copy the SKILL.md content to .claude/skills/pdf/SKILL.md
2. Ask Claude Code to work with any PDF file in your project

Example

# Extract text from a PDF (generated by Claude Code with pdf skill)
import pdfplumber

def extract_tables(pdf_path):
    tables = []
    with pdfplumber.open(pdf_path) as pdf:
        for page in pdf.pages:
            page_tables = page.extract_tables()
            for table in page_tables:
                tables.append(table)
    return tables

# Create a new PDF with reportlab
from reportlab.lib.pagesizes import letter
from reportlab.pdfgen import canvas

def create_report(output_path, title, content):
    c = canvas.Canvas(output_path, pagesize=letter)
    c.setFont('Helvetica-Bold', 16)
    c.drawString(72, 750, title)
    c.setFont('Helvetica', 12)
    y = 720
    for line in content.split('\n'):
        c.drawString(72, y, line)
        y -= 15
    c.save()

Related on TokRepo

• AI coding tools — development productivity skills
• Documentation tools — document processing resources

Common pitfalls

• Using PyPDF2 for text extraction when pdfplumber gives better results for complex layouts
• Not handling password-protected PDFs, which require the password parameter in open calls
• Attempting to extract text from scanned PDFs without OCR (use pytesseract for those cases)