Guide · Agent skill

How to use the PDF Processing Guide agent skill (hive.pdf)

Updated 2026-10-29 · written with AI from the project's own documentation and checked against our catalogue data

The hive.pdf skill, titled PDF Processing Guide, gives an AI agent a practical playbook for working with PDF files. Once it is installed, the agent knows which Python libraries and command-line tools to use for tasks such as pulling text and tables out of a PDF, merging or splitting documents, rotating pages, and OCR-ing scanned pages. This guide explains what the skill does, how to install it, how to ask for things in plain language, and what it needs to run.

PDF Processing Guideaden-hive/hive · Apache-2.0 Download the skill

Quick start

1Download the ZIP. Get the hive.pdf skill ZIP from this page.
2Unzip into a skills folder. Place the skill in ~/.claude/skills/ for personal use, or in a project's .claude/skills/ folder.
3Install the requirements. Install the Python packages with pip and make sure qpdf, poppler-utils and, for OCR, Tesseract are available.
4Make a request. Ask for the PDF task in plain language, naming the file and the result you want.
5Review the output. Open the resulting file or spreadsheet and check it, especially for important or complex documents.

What an agent skill is and what this one does

An agent skill is a folder of instructions and helper material that an AI agent can load when a task calls for it. This skill's main file, SKILL.md, is a guide to PDF work. It tells the agent how to approach common jobs and gives it ready-made snippets to run.

The skill is licensed under Apache-2.0. Its description says it covers reading, writing, merging, splitting, rotating, watermarking, encrypting and OCR-ing PDF files.

Extract text, and tables into spreadsheets
Merge, split and rotate PDFs
Add a watermark to every page
Password-protect a PDF
Extract embedded images
OCR a scanned PDF to recover text
Create new PDFs with formatted text

Install the hive.pdf skill

Download the ZIP for the skill from this page and unzip it. The skill folder contains SKILL.md and REFERENCE.md.

For a personal setup, unzip the skill into ~/.claude/skills/. To limit it to one project, unzip it into that project's .claude/skills/ folder instead.

In the Claude apps, you can instead upload the ZIP under Settings → Capabilities → Skills.

Install the requirements

The skill's examples rely on several tools. The source lists Python packages to install with pip: pypdf, pdfplumber, reportlab, pandas, pytesseract and pdf2image. OCR also needs the Tesseract OCR engine and poppler-utils, which provide the pdftotext and pdfimages commands. Command-line merging, splitting and rotating use qpdf. pdftk is optional, and pypdfium2 is referenced in REFERENCE.md for page rendering.

Python with pypdf, pdfplumber, reportlab, pandas, pytesseract and pdf2image (installed via pip)
Tesseract OCR engine and poppler-utils
qpdf command-line tool
pdftk (optional)
pypdfium2 (referenced in REFERENCE.md)

pip install pytesseract pdf2image

Trigger the skill with natural requests

You do not need special syntax. The skill is meant to be used when you ask for a PDF task in plain language. Its description lists the triggers: extracting text, tables or images from a PDF; creating or modifying a PDF; combining or splitting PDFs; OCR on a scanned PDF; password-protecting a PDF; and rendering pages as images.

Name the file and say what you want done with it. For example, you might ask the agent to combine three named reports into one PDF, or to pull the tables from page 4 of a report into an Excel file.

Merge doc1.pdf, doc2.pdf and doc3.pdf into one file
Split input.pdf so each page is its own PDF
Rotate page 1 of scan.pdf by 90 degrees
Extract all tables from report.pdf into an Excel workbook
Recover the text from this scanned PDF using OCR
Add a watermark to every page of document.pdf
Render each page of document.pdf as a PNG image

Command-line examples the agent can use

For quick jobs, the skill includes poppler-utils and qpdf commands. The page-range flags below come straight from the guide, and the agent can run similar commands on your files.

pdftotext -layout keeps the layout of the original page when extracting text
pdftotext -f 1 -l 5 extracts pages 1 to 5
qpdf --empty --pages file1.pdf file2.pdf -- merged.pdf merges files
pdfimages -j extracts images as JPEG files

pdftotext -layout input.pdf output.txt
qpdf --empty --pages file1.pdf file2.pdf -- merged.pdf
pdfimages -j input.pdf output_prefix

Safety notes: scripts and network

This skill runs scripts, so the agent executes Python and command-line code on your machine when it performs these tasks. Review what it plans to run, especially on files you care about. The source says the skill does not use the network.

The skill includes a password-removal example using qpdf --decrypt. Only use it on files you own or have permission to open. The skill is not a substitute for checking that a file is handled correctly.

Runs scripts: yes
Uses network: no
Only decrypt PDFs you own or are authorised to open

Tips for better results

Give the agent the exact file names and page numbers. Tell it whether a file is scanned, because OCR is needed for scanned pages and plain text extraction will return nothing useful.

For ReportLab documents, avoid Unicode subscript and superscript characters. The built-in fonts do not include those glyphs and they show as black boxes. Ask the agent to use the <sub> and <super> markup inside Paragraph objects instead.

Name the files and page ranges
Say when a PDF is scanned
Ask for Excel output when you want tables in a spreadsheet
Check the output on important documents

What people use it for

Batch-extract tables from PDF reports into a spreadsheet for analysis
Merge, split or rotate PDFs without learning the underlying tools
Recover text from scanned documents using OCR
Add a watermark to a set of pages or password-protect a file you own
Render PDF pages as images for use elsewhere

Frequently asked questions

Is the hive.pdf skill free?

The skill is licensed under Apache-2.0, as listed on this page. The tools it uses, such as pypdf, pdfplumber, reportlab and poppler-utils, are separate projects with their own licences, which this guide does not cover in detail.

What do I need installed before using it?

You need Python with pypdf, pdfplumber, reportlab, pandas, pytesseract and pdf2image, plus qpdf. OCR also needs the Tesseract OCR engine and poppler-utils. pdftk is optional.

How do I install an agent skill like this one?

Unzip the downloaded ZIP into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder. In the Claude apps, you can upload the ZIP under Settings → Capabilities → Skills.

Why does OCR not work on my scanned PDF?

OCR depends on the Tesseract OCR engine and poppler-utils being installed, as well as the Python packages. Check those first. The source also notes that OCR and extraction can miss or misorder content, so review the output.

Can the skill remove a password from a PDF?

The skill shows a qpdf command for decrypting a file when you supply the password. Use it only on PDFs you own or have permission to open.

How does this compare with other PDF tools?

The source does not compare this skill with other tools. It bundles Python libraries and command-line tools such as pypdf, pdfplumber, qpdf and poppler-utils, and the agent runs them for you. If you do not want the agent to run scripts or install packages, a manual tool or a service may suit you better.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Use hive.pdf if you want an agent to handle routine PDF jobs such as merging, splitting, rotating, extracting tables or OCR-ing scans, and you can install the listed tools. Pick something else if you need guaranteed accuracy on complex or legally sensitive documents, cannot install Python packages or system tools, or do not want an agent running scripts on your machine.

Download the skill