Sandbox
@ultimatile/arxiv-skills

Claude skills for arXiv lookup and paper conversion

This repository adds two Claude skills for working with arXiv papers. One skill fetches a paper, converts LaTeX or PDF into Markdown, and keeps math and section structure intact; the other searches arXiv and pulls metadata like journal DOIs.

42 stars3 forksPythonUpdated 10d ago
Who it's for

Builders who use Claude Code to search for arXiv papers, extract metadata, and turn papers into readable reference docs.

What it delivers

You can move from an arXiv ID or search query to a structured paper summary and implementation reference without doing the conversion by hand.

What it does

Paper-to-Markdown conversion

Fetches arXiv papers from source or PDF and converts them into readable Markdown.

Math and structure preservation

Keeps formulas in MathJax or LaTeX form and preserves section hierarchy.

PDF and LaTeX fallback flow

Tries LaTeX source first, then falls back to PDF, with advanced vision-based conversion for difficult papers.

ArXiv metadata lookup

Finds journal DOIs from arXiv IDs and searches papers by title, author, abstract, or category.

Claude plugin packaging

Ships as a Claude plugin bundle that can be installed from the marketplace.

How to get it

  1. 1Run
    claude plugin marketplace add ultimatile/arxiv-skills
    claude plugin install arxiv-skills
  2. 2arxiv-doc-builder can also be installed as a standalone CLI tool to run convert-paper…
    uv tool install --from 'arxiv-doc-builder @ git+https://github.com/ultimatile/arxiv-skills.git#subdirectory=skills/arxiv-doc-builder'
  3. 3Run
    # Usage
    convert-paper 2409.03108
    convert-paper 2409.03108 --output-dir ~/papers

README

arXiv Skills Repository

Custom Claude skills for working with arXiv papers - fetching, searching, and converting to reference documentation.

Available Skills

1. arxiv-doc-builder

Automatically convert arXiv papers to well-structured Markdown documentation for implementation reference.

Description: Automatically fetches arXiv papers (LaTeX source or PDF), converts them to Markdown, and generates implementation-ready reference documentation with preserved mathematics and section structure.

Use when: You need Claude to convert an arXiv paper into readable Markdown documentation for code implementation or research reference.

Capabilities:

  • Automatic paper fetching with source→PDF fallback
  • LaTeX source → Markdown conversion (via pandoc)
  • PDF → Markdown text extraction
  • Mathematical formula preservation in MathJax/LaTeX format
  • Section structure and hierarchy preservation
  • Advanced vision-based PDF conversion available for complex formulas

2. arxiv-lookup

Lightweight scripts for querying the arXiv API directly — get journal DOIs from arXiv IDs, or search for papers by title/keyword.

Description: Look up arXiv paper metadata via the arXiv API. Get journal DOIs from arXiv IDs (for OpenAlex integration), or find arXiv IDs from title/keyword search (for arxiv-doc-builder).

Use when: You need Claude to look up a paper's journal DOI, or find arXiv IDs by searching titles, authors, or categories.

Capabilities:

  • Get journal DOI from an arXiv ID (get_doi.py)
  • Search arXiv by title, author, abstract, or category (search_id.py)
  • Supports arXiv API field prefixes (ti:, au:, abs:, cat:) and boolean operators

Installation

claude plugin marketplace add ultimatile/arxiv-skills
claude plugin install arxiv-skills

Install convert-paper as a Global CLI Tool

arxiv-doc-builder can also be installed as a standalone CLI tool to run convert-paper from anywhere:

uv tool install --from 'arxiv-doc-builder @ git+https://github.com/ultimatile/arxiv-skills.git#subdirectory=skills/arxiv-doc-builder'
# Usage
convert-paper 2409.03108
convert-paper 2409.03108 --output-dir ~/papers

Requirements:

  • Python 3.11+
  • pandoc (for LaTeX conversion: brew install pandoc)
  • poppler-utils (for advanced PDF processing: brew install poppler)
  • Python dependencies auto-installed via uv

Files in the repo

Repository payload8 top-level entries
  • .claude-plugin
  • skills
  • .git-blame-ignore-revs
  • .gitignore
  • .pre-commit-config.yaml
  • LICENSE
  • README.md
  • ruff.toml

Discussion (0)

Ask about usage, or say what you built with it

Sign in to join the discussion.

No comments yet. Be the first to say what this is good for.

More skills

obra/
superpowers

An agentic skills framework & software development methodology that works.

285k
1 add

Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local deterministic AST parsing, every edge explained, no vector store.

117k
1 add
Vincentwei1021/
anything2explainer

Topic in, narrated explainer video out. A Claude Code / Codex skill that turns any topic into a black-canvas motion-graphics explainer video with TTS voiceover, subtitles and a chapter progress bar. Chinese or English; every frame drawn in code with Remotion.

666

Open-source AI job search: scan job portals, evaluate listings into a structured A-H report with a global 1-5 score, tailor your CV, track applications — runs locally in your AI coding CLI (Claude Code, Codex, OpenCode, Antigravity…)

71k