literature-overview

name	literature-overview
description	Quick landscape scan — discover papers on a topic without full-text reading
type	sop
layer	sop
agents	["alphaxiv","semantic-scholar"]
tools	{"alphaxiv":["discover_papers"],"semantic-scholar":["relevanceSearch"]}
input	query (string)
output	PaperList[] with title, authors, year, citationCount, abstract snippet
dependencies	{"sops":["literature-research","literature-search"]}

Literature Overview SOP

Layer Rules

• Layer: sop — wraps MCP tools directly
• Called by: Any tactic or strategy requiring a quick literature landscape scan
• Calls: alphaxiv MCP tools, semantic-scholar MCP tools (never calls other SOPs)

Purpose

Fast landscape scan. Understand what papers exist on a topic, who the key authors are, and rough citation counts. No full-text reading. This skill is for orientation — getting a bird's-eye view before committing to deeper reading.

Use this when you need to:

• Quickly assess how much literature exists on a topic
• Identify key papers and authors in a field
• Get citation counts to gauge paper impact
• Decide which papers deserve deeper reading (via literature-search or literature-research)

Tools

Tool	Purpose	Returns
alphaxiv.discover_papers	Semantic search for arXiv papers	Ranked paper list with title, abstract snippet, arXiv ID
ss.relevanceSearch	Keyword search across all venues	Title, abstract, authors, year, citationCount, paperId

Tool Roles

• alphaxiv.discover_papers = primary search for arXiv-covered fields (CS, math, physics, stats, EE, quant-bio/finance)
• ss.relevanceSearch = supplementary search for non-arXiv papers (biomedical, clinical, social science, humanities)

HARD-GATE

**This skill returns abstracts and metadata ONLY.**

Do NOT draw conclusions about:

• Methodology details
• Experimental results
• Specific contributions or findings
• Comparative analysis between papers

Abstracts are for ORIENTATION — identifying what exists and what looks promising.

For any substantive analysis, escalate to:

• literature-search — read AI-summarized reports (medium depth)
• literature-research — read raw full text (deep)

Treating abstracts as sufficient for research conclusions is PROHIBITED.

Workflow

Step 1: Search arXiv via alphaxiv

alphaxiv.discover_papers(
  keywords: ["keyword1", "keyword2", "keyword3"],
  question: "Detailed semantic description of desired papers",
  difficulty: 3
)