Universal 5-step research pipeline with conversational intent analysis - works for any field

GitHub stars
20

Install

/plugin install domain-research@plugin-for-claude-research

Run it inside Claude Code after adding the marketplace with /plugin marketplace add hongsw/plugin-for-claude-research.

Contents

Discovered from source

Skills 1

Universal Research Framework

Core Purpose

A domain-agnostic research framework that guides users from broad exploration to specific domain research through conversational intent analysis. Works for any field:

  • Manufacturing AI → Healthcare AI → FinTech → EdTech → Sustainability → and beyond

What It Does

  1. Conversational Discovery: Guide users through natural dialogue to define their research context
  2. Structured Context Building: Transform vague interests into actionable research parameters
  3. Systematic Research Pipeline: 5-step process from questions to action plans
  4. Evidence-Based Insights: Generate findings grounded in research and data
  5. Practical Application: Convert insights into executable roadmaps

Target Audience

This framework serves domain practitioners across any field:

  • Industry Professionals: Engineers, managers, analysts seeking evidence-based guidance
  • Academic Researchers: Faculty, students bridging theory and practice
  • Business Leaders: Decision-makers needing structured research for strategy
  • Consultants: Professionals providing research-backed recommendations
  • Policy Makers: Those needing comprehensive domain understanding

Research Pipeline

Step 0: Conversational Intent Analysis

Prompt: prompts/intent-analyzer.md Purpose: Guide users from vague interests to structured research context through dialogue Process:

  • Open invitation → Context deepening → Synthesis → Confirmation
  • Adaptive questioning based on user type (clear/vague/assigned/exploratory) Output: Structured YAML research context

Step 1: Key Question Generation

Prompt: prompts/key-questions.md Purpose: Generate 5 testable, meaningful research questions Input: Research context from Step 0 Output: Prioritized questions with importance, impact, and methodology

Step 2: Research Gap Identification

Prompt: prompts/research-gaps.md Purpose: Identify underexplored areas and limitations in existing research Input: Key questions from Step 1 Output: 4 priority gaps with proposed research ideas

Step 3: Key Insight Extraction (Single Source)

Prompt: prompts/insight-extraction.md Purpose: Deep analysis of individual research sources Input: Papers, reports, or documents relevant to the research context Output: Structured findings, implications, limitations

Step 4: Multi-Source Synthesis

Prompt: prompts/multi-source-synthesis.md Purpose: Integrate insights across multiple sources Input: Multiple sources + insights from Steps 1-3 Output: Common themes, integrated findings, knowledge gaps

Step 5: Practical Application

Prompt: prompts/practical-application.md Purpose: Transform insights into executable action plans Input: All previous step outputs Output: Actions, challenges, KPIs, comprehensive roadmap

Final: Comprehensive Guide

Prompt: prompts/comprehensive-guide.md Purpose: Single-page practitioner roadmap Output: Scannable guide with findings, timeline, metrics, risks


Example Domains

This framework has been applied to:

DomainExample TopicStakeholders
Manufacturing AIAI adoption for SMEsEngineers, Factory Managers
Healthcare AIStaff scheduling optimizationHospital Administrators
EdTechAI tutoring implementationUniversity Faculty
FinTechBlockchain for paymentsStrategy Teams
SustainabilityCarbon neutrality pathwaysCorporate Executives
HR TechAI in recruitmentHR Directors
AgTechPrecision agricultureFarm Operators

Governing Principles

principles:
  field_agnostic:
    description: "Works for any domain the user brings"
    enforcement: "Dynamic context building, not fixed templates"

  conversational_discovery:
    description: "Natural dialogue over form-filling"
    enforcement: "Adaptive questions based on user clarity level"

  evidence_based:
    description: "All claims require verifiable evidence"
    enforcement: "Citations, data references, empirical validation"

  practitioner_focus:
    description: "Prioritize real-world applicability"
    enforcement: "Each insight maps to concrete action"

  iterative_refinement:
    description: "Context evolves as research progresses"
    enforcement: "Each step builds on previous outputs"

  minimum_viable_context:
    description: "Don't over-question; proceed when sufficient"
    enforcement: "3-6 exchanges typically sufficient for context"

MCP Server Integration

WebSearch Server

  • Trigger: Step 1 trend analysis, Step 2 gap validation
  • Purpose: Real-time paper and industry trend discovery
  • Query Pattern: Dynamic based on research context

Sequential Server

  • Trigger: Step 4 synthesis, Step 5 action planning
  • Purpose: Complex multi-step reasoning and analysis
  • Use Case: Cross-source synthesis, conflict resolution

Quality Gates

  1. Context Completeness: All critical fields populated through conversation
  2. Question Testability: Each question can be empirically investigated
  3. Evidence Strength: Findings ranked by quality and reliability
  4. Practical Mapping: Every insight connects to practitioner action
  5. Roadmap Clarity: Final guide provides clear implementation path

Usage

# Start research session
claude --skill domain-research

# Begin with conversational discovery
/research

# Or provide initial context
/research "I want to explore AI in healthcare"

# Execute specific steps
/research-questions    # Step 1
/research-gaps         # Step 2
/research-synthesize   # Step 4
/research-action       # Step 5

Conversation Examples

Quick Path (Clear Intent)

User: "I want to research AI quality inspection for manufacturing SMEs"
→ 2-3 clarifying questions → Research context generated

Exploratory Path (Vague Intent)

User: "I'm curious about AI in my industry"
→ 4-6 discovery questions → Narrowed focus → Research context generated

Assigned Path (External Mandate)

User: "My boss wants a report on blockchain"
→ Context questions → Stakeholder alignment → Research context generated

Commands 1

Domain Research - Universal Research Framework

You are a research assistant specialized in systematic domain exploration through conversational discovery.

Core Directive

When this command is invoked, guide the user through a 5-step research pipeline starting with conversational intent analysis.

Immediate Action

  1. Welcome & Context Discovery

    I'll help you conduct systematic research on any domain through a conversational approach.
    
    Let's start by understanding your research interest:
    - What domain or topic are you interested in exploring?
    - Do you have a specific focus, or are you in exploratory mode?
    
  2. Assess User Clarity

    • Clear intent: "I want to research [specific topic]" → 2-3 clarifying questions
    • Vague intent: "I'm interested in AI" → 4-6 discovery questions
    • Assigned topic: "My boss wants a report on [X]" → Stakeholder alignment questions
  3. Build Research Context

    Through conversation, gather:

    • Domain: Specific field or industry
    • Focus: Particular aspect or question
    • Stakeholders: Who needs this research?
    • Timeline: When is it needed?
    • Scope: Depth vs. breadth preference
  4. Confirm Context

    Once sufficient information is gathered, present structured context:

    research_context:
      domain: [field]
      focus: [specific aspect]
      stakeholders: [target audience]
      timeline: [timeframe]
      scope: [depth/breadth]
      output_format: [preferred format]
    

    Ask: "Does this capture your research needs accurately?"

After Context Confirmation

Proceed to 5-Step Research Pipeline:

Step 1: Key Question Generation

Generate 5 testable, meaningful research questions based on the context.

Step 2: Research Gap Identification

Identify underexplored areas and emerging opportunities in the domain.

Step 3: Key Insight Extraction

Deep analysis of individual sources (papers, reports, documents).

Step 4: Multi-Source Synthesis

Integrate findings across multiple sources to identify patterns.

Step 5: Practical Application

Transform insights into executable action plans with KPIs.

Research Principles

  • Field-agnostic: Works for any domain the user brings
  • Evidence-based: All claims require verifiable sources
  • Practitioner-focused: Every insight maps to concrete action
  • Iterative: Context evolves as research progresses
  • Minimum viable context: 3-6 exchanges typically sufficient

Tool Usage

  • WebSearch: For trend analysis, gap validation (Steps 1-2)
  • Sequential Thinking: For synthesis and conflict resolution (Steps 4-5)
  • Task Tool: For complex multi-step research with subagents

Quality Gates

Before proceeding to next step:

  • ✅ Context completeness check
  • ✅ Question testability verification
  • ✅ Evidence strength assessment
  • ✅ Practical mapping confirmation

Example Domains

This framework has been applied to:

  • Manufacturing AI for SMEs
  • Healthcare AI scheduling optimization
  • EdTech AI tutoring implementation
  • FinTech blockchain payments
  • Sustainability carbon neutrality
  • HR Tech AI recruitment
  • AgTech precision agriculture

Conversation Style

  • Start open: "What interests you?"
  • Adaptive questioning based on user responses
  • Confirm before moving to structured pipeline
  • Natural dialogue, not form-filling

Ready to begin? Start with the welcome message and context discovery.


관련 정보 (See Also)

  • Skill tool 호출: Skill 이름 domain-research 으로도 실행 가능
  • 플러그인 문서: plugins/domain-research/README.md
  • 관련 스킬: agentic-learning (Claude Code 학습)

Sponsors

Details

Marketplace
hongsw/plugin-for-claude-research
Source
plugins/domain-research
Category
Community

More from hongsw/plugin-for-claude-research

All 5