arXiv Paper Search Skill

Search the arXiv repository for academic papers and preprints.

Usage

When the user asks to search for papers, use the search script:

cd /Users/varde/code/arxiv-zotero-obsidian && python3 .claude/skills/arxiv-search/scripts/search_arxiv.py --query "SEARCH_TERMS" --max-results 10

Available Options

Option Description Default

--query

Search query (required)

--max-results

Maximum results 10

--sort-by

Sort: relevance, submitted_date, last_updated relevance

--category

arXiv category (e.g., cs.AI, physics.cond-mat)

--date-from

Filter from date (YYYY-MM-DD)

--date-to

Filter until date (YYYY-MM-DD)

--output

Output format: text, json text

Query Syntax

arXiv supports field-specific searches:

• ti:keyword

• Search in title

• au:name

• Search by author

• abs:keyword

• Search in abstract

• cat:category

• Filter by category

Examples

General topic search

python3 .claude/skills/arxiv-search/scripts/search_arxiv.py --query "quantum machine learning"

Search by author

python3 .claude/skills/arxiv-search/scripts/search_arxiv.py --query "au:Hinton"

Search with category filter

python3 .claude/skills/arxiv-search/scripts/search_arxiv.py --query "deep learning" --category "cs.LG"

Combined query

python3 .claude/skills/arxiv-search/scripts/search_arxiv.py --query "ti:transformer AND abs:attention"

Recent papers only

python3 .claude/skills/arxiv-search/scripts/search_arxiv.py --query "large language models" --date-from "2024-01-01"

Output Format

The script outputs paper details including:

• arxiv_id: Unique identifier (use this for downloading)

• title: Paper title

• authors: Author names

• abstract: Paper abstract (truncated in text mode)

• published: Publication date

• pdf_url: Direct PDF download link

• categories: Subject categories

Workflow

• Run the search with user's query

• Display results in a numbered list

• Ask user which paper(s) they want to download/add to Zotero

• Note the arxiv_id for use with zotero-add skill