research-strategy

name	research-strategy
description	Conduct systematic research with confidence scoring, source validation, and structured reporting for technology decisions and codebase analysis. Use for complex research tasks, technology selection, or best practice discovery.

Research Strategy Skill

When to Activate

Activate this skill when:

• Researching technology options
• Evaluating libraries or frameworks
• Investigating best practices
• Analyzing security concerns
• Making architectural decisions
• Performing codebase analysis

Core Principles

• Research and report, don't implement
• Multiple sources beat single sources
• Document confidence levels for all findings
• Acknowledge knowledge gaps openly
• Present alternatives objectively

7-Step Research Methodology

1. Define Research Questions - What exactly needs answering?
2. Identify Information Sources - Where to look?
3. Gather Raw Information - Collect systematically
4. Cross-Reference Findings - Verify across sources
5. Validate Accuracy - Check dates, authority, consensus
6. Identify Gaps - What's still unknown?
7. Synthesize Insights - Connect into actionable knowledge

Source Prioritization

1. Primary Sources (Highest priority)

   • Official documentation
   • Technical specifications
   • API references

2. Authoritative Sources

   • Well-maintained libraries
   • Industry standards (OWASP, NIST)
   • Academic papers

3. Community Sources

   • Stack Overflow discussions
   • GitHub issues/PRs
   • Technical blogs

4. Experimental Sources (Use with caution)

   • Beta features
   • Draft proposals

Confidence Scoring

Assign to ALL findings:

Level	Range	Criteria
HIGH	90-100%	Multiple authoritative sources agree, widely adopted
MEDIUM	60-89%	Good documentation, some adoption, minor disagreements
LOW	30-59%	Limited sources, conflicting information
SPECULATIVE	<30%	Educated guess, no direct sources

Research Report Structure

# Research Report: [Topic]

## Executive Summary
[2-3 paragraphs: key findings, recommendation, confidence]

## Research Questions
1. [Question 1]
2. [Question 2]

## Key Findings

### Finding 1: [Title]
**Confidence**: HIGH/MEDIUM/LOW
**Sources**: [List with links]
[Detailed explanation]
**Implications**: [How this affects decisions]

## Comparative Analysis

| Aspect | Option A | Option B |
|--------|----------|----------|
| Performance | Details | Details |
| Learning Curve | Details | Details |

## Best Practices
1. **[Practice]**: Why, Source, Adoption level

## Risks and Concerns
- **Risk**: Severity, Likelihood, Mitigation

## Knowledge Gaps
- **Gap**: Impact, How to address

## Recommendations
[Clear, actionable recommendations with confidence levels]

Library/Framework Research Template

## Overview
- Purpose: [One sentence]
- Maturity: Stable/Beta/Experimental
- Last commit: [Date]
- License: [Type]

## Technical Assessment
- Performance: [Benchmarks]
- Bundle Size: [KB]
- Dependencies: [Count, quality]

## Developer Experience
- Documentation: Excellent/Good/Fair/Poor
- TypeScript: Built-in/DefinitelyTyped/None
- Learning Curve: Steep/Moderate/Gentle

## Verdict
**Recommendation**: Use/Don't Use/Conditional
**Confidence**: HIGH/MEDIUM/LOW
**When to Use**: [Scenarios]
**When to Avoid**: [Scenarios]

Analysis Framework

For each topic, answer:

Current State

• What exists today?
• What patterns are established?

Best Practices

• What do experts recommend?
• What anti-patterns to avoid?

Trade-offs

• What are alternatives?
• Pros/cons of each option?

Risks

• What could go wrong?
• Common pitfalls?

Anti-Patterns to Avoid

❌ Single Source Syndrome

• "According to this one article..."
• ✅ "Multiple sources agree (A, B, C)..."

❌ Premature Implementation

• "Here's the code..."
• ✅ "Implementation would follow this approach..."

❌ Missing Confidence Levels

• "This is the way."
• ✅ "HIGH confidence: Recommended by [sources]..."

❌ Outdated Information

• Using 2020 practices without checking updates
• ✅ Verify current practices, note recent changes

Research Quality Checklist

Completeness

• All questions answered
• Multiple sources consulted (minimum 2-3)
• Advantages AND disadvantages investigated
• Edge cases considered

Accuracy

• Sources are authoritative and current
• Publication dates checked
• Conflicting info acknowledged
• Assumptions stated

Actionability

• Findings translate to next steps
• Risks quantified
• Alternatives provided
• Decision criteria clear

Related Resources

See AgentUsage/research_workflow.md for complete documentation including:

• Security research template
• Detailed report examples
• Research mission template
• Quality checklists