Skills pdf
📄

pdf

Safe ⚡ Contains scripts📁 Filesystem access

Process PDF documents automatically

Also available from: 92Bilal26,92Bilal26,davila7,ArtemisAI,AutumnsGrove,Azeem-2,K-Dense-AI,DYAI2025,ComposioHQ,Cam10001110101

PDF documents require manual effort to extract data, fill forms, or combine multiple files. This skill provides programmatic tools to extract text and tables, create new PDFs, merge and split documents, and fill forms automatically using established Python libraries.

Supports: Claude Codex Code(CC)
🥉 76 Bronze
1

Download the skill ZIP

2

Upload in Claude

Go to Settings → Capabilities → Skills → Upload skill

3

Toggle on and start using

Test it

Using "pdf". Extract the text and tables from the quarterly report PDF

Expected outcome:

  • - Found 12 pages in the PDF
  • - Extracted 245 paragraphs of text
  • - Identified 3 tables with structured data
  • - Table 1: Revenue by Region (4 rows x 3 columns)
  • - Table 2: Expenses by Category (6 rows x 3 columns)
  • - Table 3: Growth Metrics (5 rows x 2 columns)

Using "pdf". Create a welcome document with company letterhead and branding

Expected outcome:

  • - Created 3-page PDF document
  • - Added company logo and letterhead on page 1
  • - Included 2 paragraphs of welcome text
  • - Added page numbers on all pages
  • - Applied brand colors to headers

Using "pdf". Fill out the tax form with the provided employee information

Expected outcome:

  • - Verified form has 15 fillable fields
  • - Filled text fields: name, address, SSN, income
  • - Set 3 checkbox fields
  • - Selected radio group option for filing status
  • - Saved completed form to completed_tax_form.pdf

Security Audit

Safe
v4 • 1/17/2026

Official Anthropic skill from github.com/anthropics/skills repository. Contains only PDF processing scripts using established Python libraries (pypdf, pdfplumber, reportlab). No network calls, no external command execution, no sensitive filesystem access beyond specified PDF files.

14
Files scanned
1,943
Lines analyzed
2
findings
4
Total audits
Audited by: claude View Audit History →

Quality Score

59
Architecture
100
Maintainability
85
Content
29
Community
100
Security
91
Spec Compliance

What You Can Build

Automate form data entry

Automatically fill PDF forms with data from other sources to reduce manual data entry work.

Extract tables from reports

Extract tabular data from PDF reports for analysis in spreadsheets or databases.

Combine case documents

Merge multiple PDF documents into organized case files with proper page ordering.

Try These Prompts

Extract text
Extract all text from input.pdf and save it to output.txt using pdfplumber.
Merge documents
Merge doc1.pdf, doc2.pdf, and doc3.pdf into merged.pdf using pypdf.
Fill form
Fill the fields in form.pdf using the data from values.json. First check if the PDF has fillable fields.
Create PDF
Create a new PDF report with title, table of quarterly sales data, and page numbers using reportlab.

Best Practices

  • Use pdfplumber for text extraction as it preserves layout better than pypdf
  • Validate form field values before filling to catch errors early
  • Convert PDFs to images first when filling non-fillable forms to ensure accurate placement

Avoid

  • Do not attempt to fill complex non-fillable forms without visual analysis
  • Do not skip bounding box validation when placing annotations
  • Do not process password-protected PDFs without obtaining the password

Frequently Asked Questions

Which PDF libraries work with Claude and Claude Code?
pypdf, pdfplumber, reportlab, pypdfium2, pdf-lib (JavaScript), and pdfjs-dist work with Claude. Each has different strengths.
What are the size limits for PDF processing?
Processing large PDFs with hundreds of pages may require chunked processing to manage memory effectively.
Can this skill integrate with existing document workflows?
Yes. Scripts can be called from any Python code. Use the scripts directory tools as building blocks for automated pipelines.
Is my data safe when processing PDFs?
All processing happens locally. Scripts read and write only specified files. No data is sent to external services.
Why is my text extraction returning empty results?
Scanned PDFs contain images, not text. Use OCR with pytesseract. If encrypted, provide the password.
How does this compare to other PDF tools?
This skill provides Claude-specific instructions. Tools like qpdf and pdftotext offer command-line alternatives.