HavenSkills

Claude PDF skills for AI agents

Find skills for extracting text and tables, running OCR on scans, editing pages, filling forms, and checking that a generated PDF renders correctly.

Browse PDF skills ↓
Skills
4
Publishers
4
First-party
3
Tested here
0

What PDF skills do

A PDF skill gives an agent a repeatable workflow for reading, editing, or producing a document. It can help the agent choose between the text layer and OCR, preserve page layout, and verify the output against the source.

  • Extract text and tables. Use the text layer for searchable PDFs. For tables, compare headers, totals and rows that span multiple pages.
  • Read scanned pages. Render scanned pages and use OCR, then spot-check names, dates and amounts before using the result.
  • Edit pages and forms. Merge, split or reorder pages, or fill an existing form and reopen the saved copy to confirm the values are visible.
  • Create and review PDFs. Generate the document, render representative pages and check for clipped content, unreadable text and awkward page breaks.

How to choose

Start with the file you have and the output you need. Select a line in your PDF viewer to see whether the file already has a usable text layer.

  • General PDF tasks Start with Anthropic PDF for extraction, OCR, page operations and interactive forms.
  • Layout-sensitive deliverables Use OpenAI PDF for creating files and visually inspecting rendered pages.
  • Advanced edits or mixed formats Review MiniMax PDF for page transformations or Documents by magnus919 if you also need Word, Excel or PowerPoint.

After processing, check page count and order, sample tables against the original, and reopen files with filled fields in a PDF viewer.

Featured PDF skills

Featured picks cover common PDF workflows. Publisher labels identify first-party sources. HavenSkills has reviewed the linked documentation and has not recorded execution tests for this directory yet.

Anthropic PDF

Extract text and tables, merge or split files, fill forms, and run OCR on scanned pages.

By Anthropic · First-party

Text extraction · Tables · OCR · Merge and split · Forms

npx skills add anthropics/skills --skill pdf

OpenAI PDF

Create PDFs with ReportLab, extract text with pdfplumber or pypdf, and inspect rendered pages before sharing.

By OpenAI · First-party

PDF creation · Visual review · Page rendering · Text extraction

More PDF skills

Explore other workflows for PDF editing and multi-format document production.

MiniMax PDF

Read PDF text and tables, run OCR, and inspect workflows for rotating, cropping, watermarking and other page edits.

By MiniMax · First-party

OCR · Page editing · Merge and split · Tables

Documents

Work with PDFs alongside Word documents, spreadsheets, and slide decks in a single document workflow.

By magnus919 · Community

PDF validation · Word · Excel · PowerPoint

npx skills add magnus919/agent-skills --skill documents

Compare PDF skills by workflow

Choose based on the work the PDF needs. For many tasks, the main decision is whether to extract data, repair pages, build a new file or validate how it looks.

How to check the result

Check a representative page and the hardest page in the document. For extracted text, compare numbers and names against the original. For a new PDF, check the rendering before sharing.

PDF skills FAQ

Which PDF skill should I install first?

Start with Anthropic PDF for general document work, including extraction, page operations and forms. Choose OpenAI PDF if visual inspection and report layout are your main concerns.

Can a PDF skill read a scanned document?

Yes, when its workflow supports OCR and the required dependencies are installed. With scans and mixed PDFs, verify the extracted text against the page image.

Do PDF skills work outside Claude Code?

Some do. Check whether your client supports the skill format and lets the agent run its required tools. The source instructions list the dependencies and any client-specific steps.

Should I install one PDF skill or an entire collection?

Start with the skill that covers your frequent job. A multi-format collection is helpful if one workflow regularly produces or reads PDFs alongside Word, Excel or PowerPoint files.

How do I ask the agent to run a PDF skill?

Name the installed skill and specify the file, page range, output and checks. For example, ask it to use the PDF skill to extract the tables on pages 3 through 5 into CSV and compare the totals with the original.

Can a skill fill every kind of PDF form?

Check whether the file has interactive fields. A flat form may need text overlays. Reopen the completed file and verify both the appearance and the saved values.