arxiv-search - SKILL.md Agent Skill

name: arxiv-search description: Use when searching arXiv papers by keyword, title, author, abstract, category, or arXiv ID through ArkSpace web_search routing.

arXiv Search

Use arXiv as a no-key web_search provider for academic paper discovery when the user asks for papers, preprints, arXiv IDs, authors, categories such as cs.AI, or literature candidates.

This skill queries the official arXiv public API and returns candidate paper metadata, abstracts, authors, categories, abstract URLs, and PDF URLs. Use a fetch or paper-reading workflow after search when the answer needs detailed synthesis from selected papers.

Source References

arXiv: https://arxiv.org/
Official arXiv API: https://info.arxiv.org/help/api/index.html
Official arXiv API user manual: https://info.arxiv.org/help/api/user-manual.html
Hermes Agent reference: https://hermes-agent.nousresearch.com/docs/user-guide/skills/bundled/research/research-arxiv

Before Use

Resolve the installed ArkSpace package root before running commands. Replace <installed-arkspace-path> with the directory two levels above this loaded SKILL.md, such as /Users/<user>/.claude/plugins/cache/ark-space/ark-space/0.1.2. Use the installed package path, not a repository-relative command.

Check readiness:

python3 <installed-arkspace-path>/scripts/arkspace.py provider check arxiv --capability web_search

arXiv search does not require provider setup or an API key.

Helper Script

Basic paper search:

python3 <installed-arkspace-path>/scripts/arkspace.py web search --provider arxiv "diffusion transformers" --max-results 5 --output json

Search by author and category:

python3 <installed-arkspace-path>/scripts/arkspace.py web search --provider arxiv "language agents" \
  --author "Yoshua Bengio" \
  --category cs.AI \
  --max-results 5 \
  --output markdown

Search by title:

python3 <installed-arkspace-path>/scripts/arkspace.py web search --provider arxiv "" \
  --title "attention is all you need" \
  --max-results 3 \
  --output json

Fetch specific IDs:

python3 <installed-arkspace-path>/scripts/arkspace.py web search --provider arxiv "" \
  --id-list 1706.03762,2402.03268 \
  --output json

Use arXiv field syntax directly when needed:

python3 <installed-arkspace-path>/scripts/arkspace.py web search --provider arxiv "ti:\"retrieval augmented generation\" AND cat:cs.CL" --max-results 10

Routing Notes

Prefer arXiv when the task is academic preprint discovery, literature scouting, paper metadata lookup, or arXiv category search.
Prefer exa-search when the task mixes papers with broader web sources, repositories, technical docs, or semantic discovery outside arXiv.
Prefer tavily-research or exa-answer when the user asks for a synthesized research report rather than a paper candidate list.
Respect arXiv API pacing for repeated paged requests; keep at least three seconds between repeated API calls.
Acknowledge arXiv data usage when publishing products or reports that materially depend on arXiv API output.