Consolidate Research

Synthesize multi-source research findings into a unified, evidence-graded report.

When to Use

Use this skill when you need to:

• •Combine research outputs from multiple AI models (Claude, Gemini, GPT)
• •Reconcile conflicting findings from different sources
• •Create a unified research synthesis with evidence scoring
• •Identify gaps and areas of consensus across sources

Prerequisites

Before using this skill, you should have:

• •Research outputs from executing a /research-brief across multiple models
• •Or multiple research documents/sources to synthesize
• •Or findings from different stakeholder interviews

Workflow

Invoke the create-research-brief skill Phase 2 for consolidation.

Input

$ARGUMENTS

Consolidation Framework

Step 1: Source Inventory

Catalog all input sources with metadata:

• •Source type (Claude, Gemini, GPT, human, document)
• •Date/recency of information
• •Credibility assessment
• •Coverage of original MECE categories

Step 2: Evidence Scoring

Apply 5-point evidence strength scale:

Step 3: Conflict Resolution (WWHTBT Protocol)

When sources disagree:

1. •What - Identify exact nature of conflict
2. •Why - Understand root cause of disagreement
3. •How - Determine resolution approach
4. •Then - Document resolution and rationale
5. •But - Note caveats and remaining uncertainty
6. •Therefore - State final consolidated finding

Step 4: Uncertainty Classification

Tag all findings with uncertainty type:

• •Aleatory - Inherent randomness (can't reduce with more data)
• •Epistemic - Knowledge gaps (can reduce with more research)
• •Model - Model limitations or biases
• •Temporal - Time-sensitive, may change

Step 5: MECE Coverage Audit

Verify consolidation covers all original categories:

• • All MECE categories addressed
• • No significant gaps remain
• • Overlaps identified and reconciled

Output Format

The consolidated research produces:

<consolidated-report>
  <header>
    <id>[unique identifier]</id>
    <sources>
      <source type="[model|document|human]" name="[source name]">
        [Source metadata]
      </source>
    </sources>
    <consolidation_date>[timestamp]</consolidation_date>
  </header>

  <executive-summary>
    [2-3 paragraph synthesis of key findings]
  </executive-summary>

  <findings-by-category>
    <category name="[MECE Category 1]">
      <finding evidence_score="[1-5]" uncertainty="[type]">
        <claim>[Consolidated finding]</claim>
        <sources>[Which sources support this]</sources>
        <conflicts>[Any disagreements and resolution]</conflicts>
      </finding>
    </category>
    <!-- ... more categories ... -->
  </findings-by-category>

  <conflicts-resolved>
    <conflict>
      <description>[Nature of conflict]</description>
      <resolution>[How it was resolved]</resolution>
      <confidence>[Confidence in resolution]</confidence>
    </conflict>
  </conflicts-resolved>

  <gaps-remaining>
    <gap priority="[high|medium|low]" type="[epistemic|data|scope]">
      [Information still needed]
    </gap>
  </gaps-remaining>

  <recommendations>
    <recommendation priority="[1-n]">
      [Actionable recommendation based on findings]
    </recommendation>
  </recommendations>

  <next-steps>
    1. [Recommended follow-up]
    2. [Additional research if needed]
  </next-steps>
</consolidated-report>

Quality Gates

• • All input sources properly inventoried
• • Evidence scores assigned to all findings
• • Conflicts explicitly identified and resolved
• • Uncertainty types classified
• • MECE coverage verified
• • Recommendations are actionable and specific
• • Executive summary captures key insights

Workflow Integration

This skill is the final step in the research pipeline:

/research-interview → /research-brief → /consolidate-research
     (elicit)            (design)           (synthesize)

After consolidation, consider:

• •Run /compare-options to evaluate recommendations
• •Run /run-expert-panel for strategic decision validation
• •Create documentation with /write-reference or /write-howto