Research Framework

Purpose

A structured deep research methodology for complex technical questions, including source evaluation, synthesis, and comprehensive reporting. Provides systematic approach to gathering, evaluating, and synthesizing information.

Problem Space:

• Ad-hoc research lacks rigor and completeness
• Source quality varies widely
• Synthesis often superficial
• Findings not actionable

Solution Approach:

• Structured research phases
• Source credibility evaluation
• Multi-perspective synthesis
• Actionable recommendations

When to Use

• Complex technical questions requiring multiple sources
• Comparative analysis (tools, frameworks, approaches)
• Investigating unfamiliar domains
• Due diligence research
• Technology evaluation
• Market/competitive analysis

When NOT to Use

• Simple factual questions with clear answers
• Questions answerable from single authoritative source
• When time constraints prevent thorough research
• Opinion-based questions without factual basis

Core Instructions

Research Framework Phases

┌─────────────────────────────────────────────────────────────────┐
│                    RESEARCH FRAMEWORK                           │
├─────────────────────────────────────────────────────────────────┤
│                                                                 │
│  Phase 1: SCOPE DEFINITION                                      │
│  ┌───────────────────────────────────────────────────────────┐ │
│  │  • Define research question precisely                      │ │
│  │  • Identify key sub-questions                              │ │
│  │  • Set scope boundaries                                    │ │
│  │  • Define success criteria                                │ │
│  └───────────────────────────────────────────────────────────┘ │
│                           │                                     │
│  Phase 2: SOURCE GATHERING                                      │
│  ┌───────────────────────────────────────────────────────────┐ │
│  │  • Primary sources (official docs, papers)                │ │
│  │  • Secondary sources (articles, tutorials)                │ │
│  │  • Community sources (forums, discussions)                │ │
│  │  • Code/implementation examples                           │ │
│  └───────────────────────────────────────────────────────────┘ │
│                           │                                     │
│  Phase 3: SOURCE EVALUATION                                     │
│  ┌───────────────────────────────────────────────────────────┐ │
│  │  • Credibility assessment                                  │ │
│  │  • Recency evaluation                                      │ │
│  │  • Bias identification                                     │ │
│  │  • Corroboration checking                                  │ │
│  └───────────────────────────────────────────────────────────┘ │
│                           │                                     │
│  Phase 4: ANALYSIS & SYNTHESIS                                  │
│  ┌───────────────────────────────────────────────────────────┐ │
│  │  • Pattern identification                                  │ │
│  │  • Contradiction resolution                               │ │
│  │  • Gap identification                                      │ │
│  │  • Insight extraction                                      │ │
│  └───────────────────────────────────────────────────────────┘ │
│                           │                                     │
│  Phase 5: REPORTING                                             │
│  ┌───────────────────────────────────────────────────────────┐ │
│  │  • Executive summary                                       │ │
│  │  • Detailed findings                                       │ │
│  │  • Recommendations                                         │ │
│  │  • Limitations & uncertainties                            │ │
│  └───────────────────────────────────────────────────────────┘ │
│                                                                 │
└─────────────────────────────────────────────────────────────────┘

Standard Procedures

Phase 1: Scope Definition

scope_definition:
  primary_question: |
    What specific question are we answering?
    (Single, clear, answerable question)

  sub_questions:
    - What background context is needed?
    - What are the key comparison dimensions?
    - What constraints apply (time, cost, technical)?
    - Who is the audience for this research?

  boundaries:
    in_scope:
      - List what IS included
    out_of_scope:
      - List what is explicitly excluded

  success_criteria:
    - What makes this research "complete"?
    - What decisions will it inform?
    - What level of confidence is needed?

Phase 2: Source Gathering

Source Categories:

Search Strategy:

1. Start broad: "[topic] overview"
2. Go deep: "[topic] architecture/internals"
3. Find comparisons: "[topic] vs [alternative]"
4. Find problems: "[topic] issues/problems/limitations"
5. Find success: "[topic] production/case study"
6. Find experts: "[topic] by [known expert]"

Phase 3: Source Evaluation

Credibility Rubric:

Red Flags:

• No author attribution
• No dates
• Sensationalist language
• No sources/citations
• Commercial bias undisclosed
• Contradicts multiple credible sources

Phase 4: Analysis & Synthesis

Synthesis Methods:

1. Thematic Analysis

   • Group findings by theme
   • Identify patterns across sources
   • Note frequency of themes

2. Comparative Matrix

   Aspect	Source A	Source B	Source C	Consensus
   Topic 1	Finding	Finding	Finding	Summary

3. Contradiction Resolution

   • When sources disagree:
     • Check recency (newer often more accurate)
     • Check authority (prefer primary sources)
     • Check context (different use cases?)
     • Note uncertainty in findings

4. Gap Analysis

   • What questions remain unanswered?
   • What would additional research reveal?
   • What assumptions are being made?

Phase 5: Reporting

Report Structure:

# Research Report: [Topic]

## Executive Summary
- Key finding 1 (confidence: high/medium/low)
- Key finding 2
- Key finding 3
- Primary recommendation

## Methodology
- Research question
- Sources consulted (count by category)
- Time period covered
- Limitations

## Detailed Findings

### Finding 1: [Topic]
**Summary**: [1-2 sentences]
**Evidence**: [Sources and data]
**Confidence**: [High/Medium/Low with justification]

### Finding 2: [Topic]
...

## Analysis

### Patterns Identified
- Pattern 1
- Pattern 2

### Contradictions & Uncertainties
- Area of disagreement 1
- Open question 1

### Gaps in Available Information
- Gap 1

## Recommendations

### Recommended Action 1
- **What**: [Specific action]
- **Why**: [Supporting evidence]
- **Risk**: [Potential downsides]
- **Confidence**: [High/Medium/Low]

## Appendix

### Sources
[Full source list with credibility scores]

### Methodology Details
[Search queries, evaluation criteria]

Decision Framework

Research Depth Selection:

When to Stop Researching:

• Saturation: New sources repeat existing findings
• Diminishing returns: Additional sources add little value
• Time/budget constraint reached
• Research question answered with sufficient confidence

Templates

Research Plan Template

research_plan:
  title: "[Research Topic]"
  date: "YYYY-MM-DD"
  researcher: "[Name]"

  question:
    primary: "[Main research question]"
    secondary:
      - "[Sub-question 1]"
      - "[Sub-question 2]"

  scope:
    in_scope:
      - "[What's included]"
    out_of_scope:
      - "[What's excluded]"
    time_period: "[Relevant date range]"

  methodology:
    depth: "quick|standard|deep|exhaustive"
    source_types:
      - primary
      - secondary
      - community
    search_queries:
      - "[Query 1]"
      - "[Query 2]"

  deliverables:
    format: "report|presentation|summary"
    audience: "[Who will use this]"
    deadline: "YYYY-MM-DD"

Source Evaluation Template

source:
  title: "[Source Title]"
  url: "[URL]"
  type: "primary|secondary|community|code|commercial"
  date_published: "YYYY-MM-DD"
  date_accessed: "YYYY-MM-DD"

  author:
    name: "[Author Name]"
    credentials: "[Relevant expertise]"
    affiliation: "[Organization]"

  evaluation:
    authority: 4  # 1-5
    accuracy: 4
    objectivity: 3
    currency: 5
    coverage: 4
    overall: 4

  key_findings:
    - "[Finding 1]"
    - "[Finding 2]"

  notes: |
    [Additional observations about this source]

  bias_concerns: |
    [Any potential biases identified]

Comparative Analysis Template

# Comparative Analysis: [Options Being Compared]

## Overview

| Criterion | Option A | Option B | Option C |
|-----------|----------|----------|----------|
| [Criterion 1] | | | |
| [Criterion 2] | | | |
| [Criterion 3] | | | |

## Detailed Comparison

### [Criterion 1]

**Option A**: [Details with sources]
**Option B**: [Details with sources]
**Option C**: [Details with sources]
**Winner**: [Option] because [reason]

### [Criterion 2]
...

## Summary

### Strengths by Option
- **Option A**: [Key strengths]
- **Option B**: [Key strengths]
- **Option C**: [Key strengths]

### Weaknesses by Option
- **Option A**: [Key weaknesses]
- **Option B**: [Key weaknesses]
- **Option C**: [Key weaknesses]

## Recommendation

**For [use case 1]**: Option A because [reason]
**For [use case 2]**: Option B because [reason]
**Default recommendation**: Option [X] because [reason]

## Confidence Level

[High/Medium/Low] - [Justification]

## Limitations

- [What this analysis doesn't cover]
- [Assumptions made]

Examples

Example 1: Framework Evaluation

Input: "Research which Python web framework to use for our new API"

Process:

1. Define scope: REST API, medium scale, team familiarity
2. Gather sources: Official docs (Django, FastAPI, Flask), benchmarks, case studies
3. Evaluate: Weight performance benchmarks higher (primary source)
4. Synthesize: Create comparison matrix on key criteria
5. Report: Recommendation with confidence levels

Output Summary:

• FastAPI recommended for greenfield API projects
• Django REST Framework for complex apps with admin needs
• Flask for maximum flexibility/minimal structure

Example 2: Technology Deep Dive

Input: "I need to understand how WebSocket scaling works"

Process:

1. Define scope: Focus on horizontal scaling, ignore single-server
2. Gather: Architecture docs, scaling guides, production case studies
3. Evaluate: Prioritize production experience reports
4. Synthesize: Identify common patterns and pitfalls
5. Report: Architecture recommendations with trade-offs

Validation Checklist

Before finalizing research:

• Primary question clearly answered
• Multiple credible sources consulted
• Source credibility evaluated
• Contradictions identified and addressed
• Confidence levels assigned to findings
• Limitations documented
• Recommendations are actionable
• Report formatted for audience

Related Resources

• Skill: technical-documentation-generator - Format research into docs
• Skill: branded-document-suite - Professional report formatting
• Critical thinking frameworks
• Academic research methodology

Changelog

1.0.0 (January 2026)

• Initial release
• Five-phase research framework
• Source evaluation rubric
• Synthesis methods
• Report templates