Skip to main content
The DNA Schema is Mega Brain’s core knowledge extraction framework. It organizes expert knowledge into 5 hierarchical layers, from foundational beliefs to actionable methodologies.

​
The 5 Layers

L1: Philosophies

Core beliefs and worldviewFundamental principles that guide all decisions and behaviors. The “why” behind everything.

L2: Mental Models

Thinking and decision frameworksWays of understanding and interpreting reality. How the expert sees the world.

L3: Heuristics

Practical rules and decision shortcutsQuick decision rules derived from experience. When to act, when to wait, when to pivot.

L4: Frameworks

Structured methodologies and processesComplete systems for solving complex problems. Repeatable approaches.

L5: Methodologies

Step-by-step implementationsDetailed execution plans with specific actions, sequences, and expected outcomes.

​
Layer Details

​
L1: Philosophies

Core beliefs that shape everything else. These rarely change and form the expert’s identity. 
filosofia:
  - id: FIL-001
    crenca: "The master is always right, except when wrong - and it's my duty to say when."
    implicacao: "Loyalty doesn't mean blind agreement. Respectful disagreement is service."
    origem: "Original JARVIS-Tony dynamic"
    confidence: 0.95
    sources: ["JV001", "JV003"]
Alex Hormozi:
  • “Price is a consequence of value” (Confidence: 0.95)
  • “Scale before you optimize” (Confidence: 0.87)
Cole Gordon:
  • “Speed beats perfection in lead response” (Confidence: 0.92)
  • “Qualification is superior to persuasion” (Confidence: 0.89)

​
L2: Mental Models

Frameworks for understanding how systems work. These shape decision-making. 
modelos_mentais:
  - id: MM-001
    modelo: "Three-Level Alert System"
    descricao: "Classify situations: Normal (continue), Attention (monitor), Critical (alert)"
    aplicacao: "Used to decide when to interrupt"
    exemplo: |
      Normal: Task proceeding, no blockers
      Attention: Dependency detected, monitoring
      Critical: Error requiring immediate decision
    sources: ["JV001"]
Alex Hormozi:
  • Value Equation: (Dream Outcome × Probability) / (Time × Effort)
  • Product Ladder: Front End → Back End → High End
JARVIS:
  • Cognitive Opportunity Cost: Every question has a cost in attention
  • Context is King: Information without context is noise

​
L3: Heuristics

Rapid decision rules and triggers. “If X, then Y” logic. 
heuristicas:
  - id: HEU-001
    regra: "When in doubt, ask. When certain, execute."
    gatilho: "Decisions with reversible vs irreversible impact"
    acao: |
      Reversible: Execute and inform
      Irreversible: Ask before acting
    sources: ["JV001", "JV002"]
Cole Gordon:
  • “Respond to leads in less than 5 minutes” (Context: Speed impacts 80% of conversion)
  • “Always anchor before revealing price” (Context: Anchoring defines value perception)
JARVIS:
  • “5 seconds of context saves 5 minutes of confusion”
  • “If it can be automated and is repetitive, it should be automated”

​
L4: Frameworks

Structured approaches to complex problems. Multi-step systems. 
frameworks:
  - id: FW-001
    nome: "SCAN - Status, Context, Action, Next"
    proposito: "Structure status updates"
    estrutura:
      - Status: "Where we are now"
      - Context: "How we got here"
      - Action: "What was done"
      - Next: "Next steps"
    exemplo: |
      Status: Pipeline phase 4 complete
      Context: Started yesterday, processed 23 files
      Action: All chunks extracted, validation OK
      Next: Begin phase 5 (narrative synthesis)
    sources: ["JV001"]
Cole Gordon:
  • SPIN Selling: Situation → Problem → Implication → Need
  • NEPQ: Neuro-Emotional Persuasion Questions framework
JARVIS:
  • SCAN: Status, Context, Action, Next (for updates)
  • DECIDE: Define, Explore, Consider, Identify, Develop, Execute

​
L5: Methodologies

Detailed step-by-step procedures. Complete execution playbooks. 
metodologias:
  - id: MET-001
    nome: "Session Initialization"
    passos:
      - passo: 1
        acao: "Load JARVIS-STATE.json"
        output: "Global state in memory"
      - passo: 2
        acao: "Check last session"
        output: "Context of where we stopped"
      - passo: 3
        acao: "Greet appropriately"
        output: "Time-appropriate greeting"
      - passo: 4
        acao: "Present status summary"
        output: "SCAN of project"
      - passo: 5
        acao: "Offer next steps"
        output: "Action suggestions"
    sources: ["JV001"]
Alex Hormozi:
  • 30-Day Onboarding Pipeline: Day 1-3 (Welcome), Day 4-7 (Quick Win), Day 8-14 (Core), Day 15-30 (Optimization)
JARVIS:
  • Session Initialization: 5-step startup sequence
  • Self-Improvement Loop: Detect → Analyze → Propose → Implement → Verify → Document

​
DNA Extraction Process

1

Insight Collection

Extract insights from source materials during Phase 4 of the pipeline
2

Layer Classification

Classify each insight into one of the 5 layers based on abstraction level
3

Confidence Scoring

Assign confidence scores (0.0-1.0) based on evidence strength
4

Source Tracing

Link every DNA element back to source chunks with chunk_id references
5

Cross-Validation

Identify contradictions, reinforcements, and consensus across sources
6

DNA Schema Generation

Compile into YAML DNA schema file for the expert

​
DNA Schema Structure

# DNA Cognitivo - Alex Hormozi
# Generated by Pipeline Jarvis v2.1

expert:
  name: "Alex Hormozi"
  id: "alex-hormozi"
  sources_count: 23
  extraction_date: "2026-01-15"

layers:
  L1_filosofias:
    - belief: "Price is a consequence of value"
      confidence: 0.95
      sources: ["AH001", "AH002"]
    - belief: "Scale before optimizing"
      confidence: 0.87
      sources: ["AH003"]

  L2_modelos_mentais:
    - model: "Value Equation"
      description: "(Dream Outcome × Probability) / (Time × Effort)"
      application: "Evaluate and improve offers"
      sources: ["AH001"]

  L3_heuristicas:
    - rule: "Never negotiate on price without anchoring value first"
      context: "Anchoring defines perception"
      sources: ["AH002"]

  L4_frameworks:
    - name: "Offer Stack Formula"
      steps:
        - "Dream Outcome identification"
        - "Probability maximization"
        - "Time/effort minimization"
      sources: ["AH001", "AH003"]

  L5_metodologias:
    - name: "Lead Magnet Launch Sequence"
      duration: "7 days"
      steps:
        - "Day 1: Tease + Problem awareness"
        - "Day 2-3: Education + Social proof"
        - "Day 4-5: Offer reveal + Scarcity"
        - "Day 6-7: Last chance + Close"
      sources: ["AH003"]

metadata:
  version: "1.0.0"
  pipeline_version: "JARVIS v2.1"
  total_chunks_analyzed: 847
  total_insights_extracted: 156

​
DNA in Agent Files

Extracted DNA schemas are used to populate agent minds:
Located in agents/minds/{expert}/DNA-CONFIG.yamlMind clones reason using the expert’s actual DNA extracted from their materials.

​
Traceability

Every DNA element traces back to source material: 
Philosophy (FIL-001)
  ↓
Sources: ["CG003", "CG007"]
  ↓
Chunks: ["CG003-042", "CG007-018"]
  ↓
Files: inbox/cole-gordon/nepq-masterclass-session-3.txt:1247
This full traceability allows you to verify any claim by jumping directly to the source material.

​
DNA Evolution

DNA schemas evolve as new material is processed:
  • New beliefs: Added with source references
  • Contradictions: Flagged for review
  • Reinforcements: Confidence scores increase
  • Refinements: Descriptions become more precise

​
Using DNA

Query DNA

Use /rag-search to query across all DNA schemas

Agent Reasoning

Mind clones use DNA to reason like the expert

Compare Experts

Theme dossiers show consensus and divergence

Build Playbooks

Compile L4/L5 into operational playbooks

​
Next Steps

Knowledge Pipeline

See how DNA is extracted in the pipeline

Agents

Learn how agents use DNA for reasoning

Build docs developers (and LLMs) love

Get started for freeTalk to us