Search arXiv, fetch paper metadata, and download PDFs into a local library.
npx skills add https://github.com/Immortalqx/my_codex_skills --skill arxiv-immortalqx
Please help me install this Agent Skill. Skill: arxiv Source: https://github.com/Immortalqx/my_codex_skills/tree/main/arxiv Command: npx skills add https://github.com/Immortalqx/my_codex_skills --skill arxiv-immortalqx
papers/ folder for your literature review. ## Quick Start Use the arxiv skill to search for recent papers on retrieval-augmented generation and download the top three PDFs into my papers folder.High-intent search queries and answers about installing and using this skill.
Run the bundled helper with a query: python scripts/arxiv_fetch.py search "attention mechanism" --max 10. It queries the arXiv Atom API and prints JSON results with IDs, titles, authors, abstracts, and PDF links.
Use the download command with the paper ID: python scripts/arxiv_fetch.py download 2301.07041 --dir papers. The script creates the directory, saves the PDF, skips files that already exist, and rejects downloads smaller than 10 KB.
Yes, both new-style IDs like 2301.07041 and legacy IDs like cs/0601001 are supported. Legacy IDs are saved with the slash replaced by an underscore, such as cs_0601001.pdf.
arXiv rate-limits requests with HTTP 429, and error pages can be returned instead of PDFs. The script retries once after 5 seconds on 429 responses and rejects any downloaded file smaller than 10 KB as a likely error page.
The API returns metadata and abstracts only, not full-text search inside papers. Results are sorted by relevance with a configurable max count, and very broad queries may need pagination via the start offset parameter.
jaredcroxtonSearch arXiv by keyword, author, category, or ID and retrieve paper metadata.
zbsayginExtract research questions, methodologies, contributions, and findings from academic papers.
xqinagVerify manuscript citations against authoritative sources with per-entry verdicts.
fcakyonSearch academic papers across arXiv, PubMed, IEEE Xplore, Scopus, ACM DL, and Semantic Scholar by keywords, authors, or topics.