Skip to content

brycewang-stanford/Auto-Empirical-Research-Skills / skills/01-lishix520-academic-paper-skills/strategist

academic-paper-strategist

Unverified

Systematic strategic planning framework for philosophy and interdisciplinary academic papers targeting preprint platforms (PhilArchive, arXiv, PhilSci-Archive). Use when users want to: (1) plan a paper on a specific topic, (2) identify research gaps and assess originality, (3) develop optimized paper outlines, (4) prepare for preprint submission, or (5) understand platform requirements and writing standards. Triggered by phrases like 'plan a paper on,' 'help me design a paper about,' 'identify research gaps in,' 'is this idea original,' or when users need structured research planning. The skill guides through three phases: Platform Analysis (identifying target venue and studying sample papers), Theoretical Framework (AI-driven literature search and gap identification), and Outline Optimization (structured design with reviewer-perspective self-assessment). Each phase includes quality evaluation standards and validation checkpoints. Output: optimized detailed outline ready for systematic writing (use with academic-paper-composer skill).

FileKindSize
SKILL.mdskill md22.3 KB
references/quality_standards.mdmarkdown10.2 KB
references/search_strategy.mdmarkdown14.0 KB

references/search_strategy.md

View on GitHub
<!--
  ╔══════════════════════════════════════════════════════════════╗
  ║  本文件为开源 Skill 原始文档,收录仅供学习与研究参考        ║
  ║  CoPaper.AI 收集整理 | https://copaper.ai                  ║
  ╚══════════════════════════════════════════════════════════════╝

  来源仓库: https://github.com/lishix520/academic-paper-skills
  项目名称: academic-paper-skills
  开源协议: MIT License
  收录日期: 2026-04-02

  声明: 本文件版权归原作者所有。此处收录旨在为社会科学实证研究者
  提供 AI Agent Skills 的集中参考。如有侵权,请联系删除。
-->

# Literature Search Strategy

This document provides systematic strategies for comprehensive literature search and gap identification.

## Overview

Effective literature search for academic papers requires:
1. **Multi-dimensional search** (keywords, time, diversity)
2. **Iterative refinement** (3 rounds minimum)
3. **Quality filtering** (relevance + authority)
4. **Gap mapping** (identify white spaces)

---

## Phase 1: Keyword Generation

### Step 1: Extract Core Concepts
From user's research idea, identify:
- **Primary concepts** (2-3): The main theoretical constructs
- **Secondary concepts** (2-3): Supporting or related ideas
- **Methodological terms** (1-2): Approach or framework

**Example**:
- Research idea: "Self-continuity through narrative compression during sleep"
- Primary: self-continuity, narrative compression
- Secondary: sleep, memory consolidation, personal identity
- Method: phenomenological analysis, computational metaphor

### Step 2: Generate Keyword Combinations
Create 10-15 search strings combining:
- Concept + concept
- Concept + method
- Concept + domain

**Example combinations**:
1. "self-continuity" AND "sleep"
2. "narrative compression" AND "memory"
3. "personal identity" AND "sleep-wake cycle"
4. "phenomenological self" AND "consciousness"
5. "autobiographical memory" AND "consolidation"
...

### Step 3: Include Synonyms and Variants
For each core concept, add:
- Disciplinary variants (philosophical vs. psychological terms)
- Historical terms (older vs. current terminology)
- Related concepts (broader/narrower)

**Example**:
- self-continuity → personal identity, diachronic self, narrative self
- sleep → periodic disruption, consciousness gaps, sleep-wake cycles
- compression → consolidation, integration, synthesis

---

## Phase 2: Multi-Round Search

### Round 1: Direct Search (Primary Literature)

**Goal**: Find 30-50 directly relevant papers

**Search Tools**:
- Use Exa MCP for semantic search
- Use Tavily MCP for web search
- Search Google Scholar / PhilPapers / arXiv as needed

**Process**:
```markdown
For each of 10-15 keyword combinations:
1. Run search with time filter (last 5 years)
2. Collect top 5 results per query
3. Total: 50-75 candidate papers

Quality filter:
- Read abstracts
- Check relevance (must score ≥7/10)
- Retain top 30-50 papers
```

**Relevance Scoring Guide** (0-10):
- 10: Exact topic match, same methodology
- 9: Topic match, different angle
- 8: Core concept match, different context
- 7: Related concept, applicable methods
- 6: Adjacent field, partial relevance
- 5: Tangential relevance
- <5: Exclude

### Round 2: Expanded Search (Adjacent Literature)

**Goal**: Find 10-20 papers from adjacent fields

**Process**:
```markdown
1. Extract new keywords from Round 1 papers:
   - Check "keywords" sections
   - Note frequently cited concepts
   - Identify theoretical frameworks used

2. Search adjacent fields:
   - If philosophy → check cognitive science
   - If neuroscience → check philosophy of mind
   - If AI → check consciousness studies

3. Collect 10-20 bridging papers
```

**Why Expanded Search**:
- Discover interdisciplinary connections
- Find alternative approaches to similar problems
- Identify potential theoretical resources

### Round 3: Classic Literature (Foundational Works)

**Goal**: Identify 5-10 highly-cited foundational papers

**Process**:
```markdown
1. Identify classics from Round 1-2 citations:
   - Look for papers cited in >50% of collected papers
   - Note "seminal works" mentioned

2. Citation tracking:
   - Use Google Scholar "Cited by" counts
   - Target papers with >100 citations
   - Focus on works from 1990-2015 (established but relevant)

3. Verify foundational status:
   - Check if it introduced key concepts
   - Assess if it shaped the field
   - Confirm current relevance
```

**Selection Criteria**:
- Citation count >100 (adjust for field size)
- Still cited in recent papers (2020+)
- Introduced terminology/concepts still in use

---

## Phase 3: Gap Identification

### Method 1: Concept Mapping

**Process**:
```markdown
1. Create 2D concept map:
   - X-axis: Concepts (from your keywords)
   - Y-axis: Methods/Approaches

2. Plot each collected paper on the map

3. Identify white spaces:
   - Concept combinations not explored
   - Methods not applied to certain concepts
   - Questions raised but not answered
```

**Example Map**:
```
              | Phenomenology | Cognitive Science | Computational |
--------------|---------------|-------------------|---------------|
Self          |   ████████    |     ████         |      ██       |
Sleep         |   ████        |     ██████       |      █        |
Compression   |   ██          |     ███          |               | ← GAP!
```

### Method 2: Problem-Solution Analysis

**Process**:
```markdown
1. List problems addressed in literature:
   - What questions do existing papers ask?
   - What puzzles do they try to solve?

2. List solutions proposed:
   - What answers/theories do they offer?
   - What limitations do they acknowledge?

3. Identify unsolved problems:
   - Problems mentioned but not addressed
   - Solutions that create new problems
   - Acknowledged limitations
```

**Gap Types**:
- **Empty gaps**: Problem identified but no solutions attempted
- **Partial gaps**: Solutions proposed but incomplete/unsatisfactory
- **Controversy gaps**: Competing solutions, no consensus

### Method 3: Temporal Gap Analysis

**Process**:
```markdown
1. Plot papers by publication year

2. Identify research waves:
   - When did interest surge?
   - When did it decline?
   - Recent vs. historical focus?

3. Spot temporal gaps:
   - Topics once active but abandoned
   - Emerging questions (last 2 years)
   - Due for re-examination with new methods
```

**Indicators of Good Temporal Gaps**:
- Last major work >5 years ago (ripe for update)
- New methods available (AI, neuroscience tools)
- Recent empirical findings enabling new theories

---

## Phase 4: Originality Assessment

### Step 1: Similarity Matrix

Create table comparing user's idea with top 10 most similar papers:

```markdown
| Paper | Topic Overlap | Method Overlap | Conclusion Overlap | Overall Similarity |
|-------|---------------|----------------|--------------------|--------------------|
| A     | 90%          | 70%           | 60%               | 73% (High)         |
| B     | 80%          | 50%           | 40%               | 57% (Medium)       |
| C     | 70%          | 80%           | 30%               | 60% (Medium)       |
...
```

**Interpretation**:
- >80%: Too similar, needs repositioning
- 50-80%: Moderate overlap, emphasize differences
- <50%: Good originality

### Step 2: Innovation Type Classification

Identify which type(s) of innovation apply:

1. **Methodological Innovation**:
   - Applying new method to known problem
   - Example: Using predictive coding to explain self-continuity

2. **Theoretical Innovation**:
   - New framework or model
   - Example: Compression metaphor for narrative self

3. **Application Innovation**:
   - Existing theory to new domain
   - Example: Memory consolidation theory applied to philosophy of self

4. **Integrative Innovation**:
   - Synthesizing previously separate literatures
   - Example: Combining phenomenology + cognitive neuroscience

**Requirement**: Must identify at least 2 innovation types for sufficient originality

### Step 3: Impact Prediction

**Scoring Criteria** (1-10 scale):

**Gap Importance** (5 points):
- 5: Core unsolved problem in the field
- 4: Significant but not central problem
- 3: Interesting niche problem
- 2: Minor technical issue
- 1: Trivial question

**Generalizability** (3 points):
- 3: Widely applicable across domains
- 2: Applicable within subdiscipline
- 1: Narrow case-specific

**Explanatory Power** (2 points):
- 2: Resolves multiple existing puzzles
- 1: Clarifies one specific issue
- 0: Purely descriptive

**Total Score Interpretation**:
- 9-10: High impact potential (major contribution)
- 7-8: Good impact (solid contribution)
- 5-6: Moderate impact (acceptable)
- <5: Low impact (reconsider direction)

---

## Phase 5: Evidence Documentation

### Gap Evidence Package

For each identified gap, create evidence file:

```markdown
## Gap: [Title]

### Definition
[Clear description of what's missing]

### Evidence from Literature
1. Paper A (Author, Year): "[Quote showing gap]"
   - Context: [Why this matters]

2. Paper B (Author, Year): "[Quote acknowledging limitation]"
   - Context: [What they couldn't address]

3. Paper C (Author, Year): "[Quote raising question]"
   - Context: [Unanswered question]

[Minimum 3 pieces of evidence per gap]

### Why This Gap Matters
[Significance explanation]

### Feasibility for User
[Can user address this with their approach?]
```

---

## Search Quality Checklist

Before concluding literature search, verify:

**Coverage**:
- ✓ All core concepts covered (each has 5-10 papers)
- ✓ Time span adequate (5 years + classics)
- ✓ Multiple disciplines searched (if applicable)
- ✓ Total papers ≥20 (minimum)

**Quality**:
- ✓ Average relevance ≥7.5/10
- ✓ At least 5 high-citation papers (>100 cites)
- ✓ Recent papers (last 2 years) included
- ✓ Foundational works included

**Gap Identification**:
- ✓ At least 3 distinct gaps identified
- ✓ Each gap has 3+ pieces of evidence
- ✓ Gaps are feasible to address
- ✓ Gaps are significant (not trivial)

**Originality**:
- ✓ Similarity analysis completed (top 10 papers)
- ✓ Innovation types identified (≥2)
- ✓ Impact score calculated (≥7/10)
- ✓ Justification documented (300+ words)

---

## Common Pitfalls and Solutions

### Pitfall 1: Search Too Narrow
**Symptom**: <15 relevant papers found
**Solution**:
- Broaden keywords (use synonyms)
- Extend time range
- Include adjacent disciplines

### Pitfall 2: Search Too Broad
**Symptom**: >100 papers, can't filter effectively
**Solution**:
- Tighten relevance criteria (raise threshold to 8/10)
- Focus on most recent papers
- Prioritize by citation count

### Pitfall 3: Missing Classics
**Symptom**: No papers >10 years old
**Solution**:
- Check "most cited" on Google Scholar
- Look at textbook references
- Ask: "What are the foundational works in this area?"

### Pitfall 4: Echo Chamber
**Symptom**: All papers from same research group/school
**Solution**:
- Deliberately search for alternative perspectives
- Check papers that cite different theoretical frameworks
- Include critics of main approach

### Pitfall 5: False Gaps
**Symptom**: "Gap" is actually well-addressed, just using different terminology
**Solution**:
- Search with synonym keywords
- Check related disciplines
- Verify gap with evidence (3+ citations)

---

## Integration with Skill Workflow

This search strategy supports:

**Phase 2.1** (Gap Identification):
- Use Rounds 1-3 to build literature base
- Apply Phase 3 methods to identify gaps
- Document evidence per Phase 5

**Phase 2.2** (Originality Assessment):
- Use Phase 4 similarity analysis
- Classify innovation types
- Calculate impact prediction

**Quality Gates**:
- Gate 2 requires passing the Search Quality Checklist
- All gaps must have evidence packages
- Originality assessment must be documented

---

## Example: Complete Search for "Self-Continuity and Sleep"

### Round 1: Direct Search
**Keywords used**: 15 combinations (self-continuity + sleep, narrative self + memory, etc.)
**Papers found**: 45 candidates
**After filtering**: 25 papers retained (relevance ≥7/10)

### Round 2: Expanded Search
**New keywords**: phenomenological self, memory consolidation, sleep-dependent processing
**Adjacent fields**: Cognitive neuroscience, philosophy of mind
**Papers added**: 12 papers

### Round 3: Classics
**Identified works**: Thompson (2015), Dennett (1991), Metzinger (2003), Born & Wilhelm (2012)
**Papers added**: 8 classic papers

**Total**: 45 papers in literature base

### Gap Identification
**Method**: Concept mapping + Problem-solution analysis

**Gaps found**:
1. **Compression mechanism**: How exactly is narrative compressed during sleep?
   - Evidence: Thompson mentions "taken apart and put together" but no mechanism
   - 5 papers acknowledge this limitation

2. **Functional explanation**: Why is self-continuity maintained (evolutionary purpose)?
   - Evidence: Most papers describe but don't explain function
   - 4 papers raise this as future work

3. **Cross-sleep identity**: Philosophical implications of discontinuity
   - Evidence: Neuroscience describes process, philosophy discusses identity, but no integration
   - Gap at intersection of two literatures

### Originality Assessment
**Similarity analysis**: Highest overlap 65% (medium)
**Innovation types**: Integrative (philosophy + neuroscience) + Theoretical (compression framework)
**Impact score**: 8/10 (bridges two literatures, addresses acknowledged gaps)

---

## Summary

Effective literature search requires:
1. **Systematic keyword generation** (10-15 combinations)
2. **Multi-round search** (direct → expanded → classics)
3. **Rigorous gap identification** (3 methods, evidence-based)
4. **Thorough originality assessment** (similarity + innovation + impact)
5. **Quality verification** (pass all checklist items)

This ensures comprehensive coverage and confident identification of genuine research opportunities.
brainstormingobra/superpowersYou MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation).Stars: 323kUpdated Claude CodeCodexVerifiedInstall
brand-guidelinesanthropics/skillsApplies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
frontend-designanthropics/skillsGuidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
mcp-builderanthropics/skillsGuide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).Stars: 198kUpdated Claude CodeCodexVerifiedInstall