systematic-literature-review

Systematic Literature Review

Overview

Produces a structured systematic literature review (SLR) across multiple academic papers on a research topic. Given a topic query, searches arXiv, extracts structured metadata from each paper, synthesizes themes, and emits a final report with consistent citations.

Distinct from academic-paper-review: that skill does deep peer review of a single paper. This skill does breadth-first synthesis across many papers.

Poirot note: The original deer-flow skill uses a bundled scripts/arxiv_search.py + subagent task tool for parallel extraction. Poirot has neither, so this version uses bash with curl to the arXiv API directly + sequential single-agent extraction.

When to Use

• A literature survey on a topic ("survey transformer attention variants")
• A synthesis across multiple papers ("what do recent papers say about X")
• A systematic review with consistent citation format
• An annotated bibliography on a topic
• An overview of research trends in a field over a time window

Do not use when:

• User provides exactly one paper (use academic-paper-review)
• User asks a factual question (answer directly)

Workflow

Phase 1: Plan

Confirm with the user:

• Topic: the research area in plain English
• Scope: how many papers (default 20, hard upper bound 50), optional time window, optional arXiv category (e.g. cs.CL)
• Citation format: APA, IEEE, or BibTeX (default APA)

If user says "50+ papers", cap at 50 and explain synthesis quality degrades past that.

Phase 2: Search arXiv

Use bash with curl to the arXiv API. Extract 2-3 core keywords before searching — don't pass the full topic description as the query.

# Search arXiv (use 2-3 core keywords, not the full topic)
curl -s "https://export.arxiv.org/api/query?search_query=all:transformer+attention&max_results=20&sortBy=relevance" | python3 -c "
import sys, xml.etree.ElementTree as ET, json
ns = {'a': 'http://www.w3.org/2005/Atom'}
root = ET.fromstring(sys.stdin.read())
papers = []
for entry in root.findall('a:entry', ns):
    papers.append({
        'id': entry.find('a:id', ns).text.split('/')[-1],
        'title': entry.find('a:title', ns).text.strip().replace('\n', ' '),
        'authors': [a.find('a:name', ns).text for a in entry.findall('a:author', ns)],
        'published': entry.find('a:published', ns).text[:10],
        'abstract': entry.find('a:summary', ns).text.strip(),
        'pdf_url': [l.get('href') for l in entry.findall('a:link', ns) if l.get('title') == 'pdf'],
        'abs_url': entry.find('a:id', ns).text,
    })
print(json.dumps(papers, indent=2, ensure_ascii=False))
"