Skip to main content

66s Review

Benchmarked against: Anthropic — Using the Evaluation Tool Skill: /66s-review (on-demand, L3 Progressive Disclosure) Framework: Six Thinking Hats (de Bono) + Six Action Shoes (de Bono) + Systems Thinking

The 66s Review is SuperPortia's systematic review framework for plans, specs, architecture decisions, and any proposal before Captain approval. It ensures decisions are examined from multiple perspectives before commitment.


When to use

TriggerExample
Major spec or planMTAAA v1.4 spec review
Architecture decisionChoosing between LangGraph vs n8n
Before HITL presentationPresenting options to Captain
WO with high blast radiusInfrastructure migration
New capability designPAM (Project Agentic Manager) design

Rule of thumb: If a decision affects multiple agents, ships, or is hard to reverse — run a 66s Review first.


The framework

Six Thinking Hats

Each "hat" represents a thinking perspective. The review cycles through all six:

HatColorPerspectiveKey question
WhiteInformationWhat data do we have? What's missing?"What are the facts?"
🟥 RedEmotion/IntuitionGut feeling — what feels right or wrong?"What does instinct say?"
BlackCautionRisks, dangers, weaknesses"What could go wrong?"
🟨 YellowOptimismBenefits, opportunities, value"What's the best case?"
🟩 GreenCreativityAlternatives, new ideas"What else could we do?"
🟦 BlueProcessMeta-thinking — is the review process working?"Are we thinking correctly?"

Six Action Shoes

Each "shoe" represents an action mode — what to do after thinking:

ShoeColorAction modeWhen to use
🔵 NavyFormal/RoutineFollow established proceduresStandard operations, SOPs
🟤 BrownPragmaticGet practical resultsImplementation, delivery
🟠 OrangeEmergencyAct quickly on dangerIncidents, critical failures
🩷 PinkCaringConsider human impactTeam dynamics, user experience
🟣 PurpleAuthorityExercise judgmentCaptain decisions, policy changes
🔘 GreyInvestigationGather more informationUnknowns, research needed

Systems Thinking

The third lens — how this decision interacts with the whole system:

CheckQuestion
DependenciesWhat depends on this? What does this depend on?
Feedback loopsWill this create positive or negative spirals?
Side effectsWhat unintended consequences are possible?
ReversibilityHow hard is it to undo?
Fleet impactDoes this affect SS1/SS2/SS3 differently?

Review output format

## 66s Review: [Subject]

### White Hat — Facts
- [Data points, current state, metrics]

### Red Hat — Intuition
- [Gut feelings, team sentiment]

### Black Hat — Risks
- [Potential failures, weaknesses, costs]

### Yellow Hat — Benefits
- [Advantages, opportunities, value]

### Green Hat — Alternatives
- [Other approaches considered]

### Blue Hat — Process
- [Is this review thorough? What's missing?]

### Action Shoes
- **Recommended shoe:** [which action mode]
- **Rationale:** [why this mode]

### Systems Thinking
- **Dependencies:** [what's affected]
- **Reversibility:** [easy/moderate/hard]

### Verdict
- **Recommendation:** [approve / approve with conditions / reject / needs more research]
- **Conditions:** [if applicable]
- **Key risks to monitor:** [top 2-3 risks]

Review scoring

Each hat dimension gets a score:

ScoreMeaning
ClearThis dimension is well-addressed
⚠️ ConcernIssues identified but manageable
🔴 BlockerCritical issue — must be resolved before proceeding

A review with any 🔴 Blocker cannot be approved until the blocker is resolved.


Integration with WO system

66s Reviews are typically triggered during the Architect phase of the ADLC:

Review results are ingested to UB with tags: 66s-review, [project], [subject].


Example reviews

SubjectVerdictKey insightUB ref
PAM Design Spec v1.1Approved with 4 recommendationsPurple shoe — this is a Captain-level architecture decisionub-7a4a003357d5
Cloud UB MigrationApprovedOrange shoe — current split is causing data loss riskub-396f44b70763
MTAAA v1.4ApprovedGreen hat found Amazon Books as taxonomy sourceMTAAA spec

PageRelationship
EGS Compliance66s Review is part of EGS compliance for major decisions
Pre-Flight CheckQuick risk check (30 seconds) vs full review (66s)
EGS SpecADLC Architect phase requires 66s Review
HITL/HOTLReview results inform Captain decisions