paper-compare

name	paper-compare
slug	paper-compare
triggers	/paper-compare
author	ClawHub community
license	Community-vendored
hub_pack	clawhub-research
trust_level	community
upstream_url	https://clawhub.ai/skills/paper-compare
audit_status	reviewed
description	Compare academic research papers side-by-side to identify similarities, differences, and research gaps. Use when user wants to compare 1-5 papers via DOIs, URLs, search queries, or PDF files. Supports mixed input types. Outputs both comparison table and detailed narrative summary.
version	1.2.1
changelog	Added user_goal to history, clarified step structure
metadata	{"clawdbot":{"emoji":"📄","category":"research"}}

Paper Compare

Compare academic papers side-by-side with structured tables and detailed narrative analysis.

---

The Paper Comparison Reasoning Framework

┌─────────────────────────────────────────────────────────────┐
│  PAPER COMPARISON THINKING                                   │
├─────────────────────────────────────────────────────────────┤
│  1. INTERPRET  → What papers? What comparison goal?        │
│  2. RETRIEVE   → Fetch metadata, abstracts, full text     │
│  3. ANALYZE    → Extract across 10 dimensions              │
│  4. SYNTHESIZE → Build narrative, find gaps, score quality │
│  5. VALIDATE   → Check completeness, deliver              │
└─────────────────────────────────────────────────────────────┘

---

Decision Tree: Input Processing

USER INPUT
    │
    ├── 1 paper ──→ Single Paper Summary
    │       └── Skip comparison, show full summary
    │
    ├── 2-5 papers ──→ Full Comparison
    │       └── Proceed with 10 dimensions
    │
    ├── >5 papers ──→ Ask to Narrow
    │       └── "Please narrow to 2-5 for meaningful comparison"
    │
    ├── DOI ──→ Fetch via crossref/semantic scholar
    │       └── https://api.crossref.org/works/{doi}
    │
    ├── URL ──→ Fetch via web_fetch
    │       └── Extract title, authors, abstract
    │
    ├── Search query ──→ Search first
    │       └── Use web_search, present top 3, CONFIRM before proceeding
    │
    └── PDF file ──→ Extract text first
            └── Use pdf skill, then extract metadata

---

Decision Tree: Comparison Angle

WHAT IS THE COMPARISON ABOUT?
    │
    ├── Same topic, different methods ──→ 
    │       └── Focus: methodology differences, results comparison
    │
    ├── Same method, different domains ──→
    │       └── Focus: adaptation, performance across domains
    │
    ├── Evolution over time ──→
    │       └── Focus: improvements, what changed, SOTA progression
    │
    ├── Competing approaches ──→
    │       └── Focus: trade-offs, when to choose which
    │
    └── Complementary papers ──→
            └── Focus: how they combine, gaps each fills

Self-Check: After Identifying Angle

• Does my analysis focus on the right aspects?
• Will this help the user make a decision?

---

Step 1: Interpret the Request

What to Clarify

Question	Why It Matters
Which papers?	Need exact references
What goal?	Learning? Research? Writing?