arxiv-literature

name	arxiv-literature
description	Use when the user asks to search arXiv, summarize an arXiv paper, compare papers, find recent research, or build a compact literature survey. Prefer the Researcher shadow agent for noisy multi-paper retrieval; return only citation-ready, source-linked paper cards.
version	0.1.0
license	Apache-2.0
allowed-tools	["web_fetch","dispatch","read","grep","find","ls"]
registry-id	iowarp/clio-coder
source-url	https://github.com/iowarp/clio-coder/tree/main/skills/arxiv-literature
audit	pass

ArXiv Literature

Find, summarize, or compare academic papers while protecting the main context window.

Prefer dispatching the researcher shadow agent for searches, comparisons, and multi-paper synthesis. Use direct web_fetch for a single known paper URL, for a small arXiv Atom query, or when dispatch is unavailable.

Route

• Single paper: arXiv URL/ID, “summarize this paper”, “explain this paper”
• Search: “find papers about X”, “search arxiv for X”, “latest papers on X”
• Compare: multiple paper URLs/IDs, “compare these papers”
• Survey: “what is the literature on X”, “best papers for X”

Preferred workflow

1. Delegate noisy retrieval

Ask the researcher shadow agent for a compact result:

Research arXiv literature for: <user goal>.
Return only compact source-linked paper cards, comparison/synthesis, caveats, and read/skim/skip recommendations.

2. If doing it directly

For paper URLs, call web_fetch on the arXiv URL. Clio normalizes arXiv paper pages into structured metadata plus AlphaXiv enrichment when available.

For search, use arXiv Atom API through web_fetch; Clio compacts the XML into paper cards:

https://export.arxiv.org/api/query?search_query=all:QUERY&sortBy=submittedDate&sortOrder=descending&start=0&max_results=10

Useful categories:

• cs.AI artificial intelligence
• cs.LG machine learning
• cs.CL NLP/LLMs
• cs.CR security
• cs.SE software engineering
• cs.MA multi-agent systems
• cs.IR information retrieval/RAG
• cs.CV vision
• cs.RO robotics

3. Return compressed output

Do not paste raw Atom XML or entire paper text. Return:

## Literature Result

### Query
...

### Best Matches / Papers
1. **Title** — authors, date, `arxiv:id`
   - Problem:
   - Method:
   - Evidence:
   - Limitation:
   - Relevance:
   - Links:

### Recommendation
- Read:
- Skim:
- Skip:

Gotchas

• arXiv Atom is XML, not JSON; use web_fetch so Clio compacts it.
• lastUpdatedDate can surface old papers with small edits. Use submittedDate for newly submitted work.
• AlphaXiv is AI-generated enrichment; useful for scanning but not authoritative.
• Fetch/enrich only top candidates. Keep the main context small.