pdf-tools
About
The pdf-tools skill handles PDF conversion, processing, and analysis, primarily converting PDF pages to JPG images for vision analysis. It provides a dedicated Python script with options for DPI settings and page ranges, enforcing organized output and two-pass optimization. Use this skill when you need to prepare PDF content for tasks like OCR or visual inspection.
Quick Install
Claude Code
Recommendednpx skills add karstenheld3/SharePoint-GPT-Middleware -a claude-code/plugin add https://github.com/karstenheld3/SharePoint-GPT-Middlewaregit clone https://github.com/karstenheld3/SharePoint-GPT-Middleware.git ~/.claude/skills/pdf-toolsCopy and paste this command in Claude Code to install this skill
GitHub Repository
Related Skills
release-standards
DocumentsThis skill provides semantic versioning (semver) guidelines and changelog formatting standards for software releases. Use it when preparing releases to correctly increment version numbers (major/minor/patch) and structure changelog entries. It includes rules for pre-release identifiers and clear examples for developers.
commit-standards
DocumentsThis skill formats Git commit messages according to the Conventional Commits standard. It provides templates and type definitions (like `feat`, `fix`, `refactor`) to ensure consistency when writing or reviewing commits. Use it during the commit process to create clear, structured commit history.
huggingface-tokenizers
DocumentsThis skill provides high-performance tokenization using HuggingFace's Rust-based library, processing 1GB of text in under 20 seconds. It supports BPE, WordPiece, and Unigram algorithms while enabling custom tokenizer training and alignment tracking. Use it when you need production-fast tokenization or to build custom tokenizers integrated with the transformers ecosystem.
nano-pdf
Documentsnano-pdf is a CLI tool that lets developers edit PDFs using natural-language instructions, like changing text or fixing typos on specific pages. It's ideal for quick, programmatic PDF modifications directly from the terminal. Always verify the output, as page numbering can vary between versions.