arXiv Search

arXiv Search gives your Claude agent direct access to over 2 million open-access academic papers across physics, mathematics, computer science, quantitative biology, statistics, and economics. No API key required — arXiv provides a public API that this skill wraps with a clean, agent-friendly interface.

Search by keyword, author name, paper ID, or category. Filter by date range to find the latest research or anchor your query to a specific period. Results include title, authors, abstract, submission date, categories, and a direct PDF download URL. The skill also fetches the full paper text when requested, chunking long documents into manageable segments for analysis.

Practical use cases abound: tracking the state of the art in machine learning, finding cited papers referenced in a bibliography, monitoring a specific research group's output, or building literature review summaries. The skill integrates naturally with Summarize to condense abstracts and with Obsidian or Notion to build a personal research library. For systematic literature reviews, chain it with the Deep Research Agent — arXiv Search handles domain-specific academic sources while Tavily covers the broader web. Academics, AI engineers, and science journalists use this skill to stay current without manually trawling arXiv's web interface daily.

Installation

clawhub install arxiv-search

Install: clawhub install arxiv-search