Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
A direct command skips the review prompt. Inspect the source before running it.
Orchestrates 9-persona CEO Board deliberation for strategic business decisions, with persistent agent expertise files
CEO Board Deliberation Skill
Overview
The CEO Board Deliberation skill orchestrates multi-agent reasoning for strategic business decisions. It implements IndyDevDan's innovation of templated agent expertise with persistent learning.
Created: 2026-03-24
Framework: CleanExpo/NodeJS-Starter-V1
Agents: 9 specialized board members with persistent expertise
The Golden Example
To understand how the board works, review this complete example decision:
Perspective: Holistic business lens, execution reality
Questions: Can we execute this? Does it align with strategy? How do we communicate it?
Expertise File:.pi/ceo-agents/expertise/ceo.md
2. Revenue Agent
Perspective: Growth, unit economics, expansion
Questions: How does this impact revenue? Is it capital-efficient? What's the market opportunity?
Expertise File:.pi/ceo-agents/expertise/revenue.md
3. Product Strategist
Perspective: Feature prioritization, user value, roadmap
Questions: Does this delight users? How does it fit the roadmap? What's the UX impact?
Expertise File:.pi/ceo-agents/expertise/product-strategist.md
4. Technical Architect
Perspective: System design, feasibility, technical debt
Questions: Is this technically sound? What's the implementation complexity? Scalability concerns?
Expertise File:.pi/ceo-agents/expertise/technical-architect.md
5. Contrarian
Perspective: Risk identification, blindspots, second-order effects
Questions: What could go wrong? What are we not seeing? When should we say "no"?
Expertise File:.pi/ceo-agents/expertise/contrarian.md
6. Compounder
Perspective: Long-term leverage, compound effects, platform thinking
Questions: What compound effects does this create? Will this give us an unfair advantage over time?
Expertise File:.pi/ceo-agents/expertise/compounder.md
7. Custom Oracle
Perspective: Unite-Group domain expertise โ 7 businesses, MACAS advisory, Brisbane/QLD AUS context
Questions: What do domain experts know about this? Regulatory/compliance concerns (Privacy Act 1988, ATO)?
Expertise File:.pi/ceo-agents/expertise/custom-oracle.md
8. Market Strategist
Perspective: Competitive positioning, market timing, go-to-market
Questions: How does this position us competitively? Is timing right? Market adoption risk?
Expertise File:.pi/ceo-agents/expertise/market-strategist.md
9. Moonshot
Perspective: Radical innovation, asymmetric bets, breakthrough possibilities
Questions: Is there a bigger opportunity here? What's the moonshot play?
Expertise File:.pi/ceo-agents/expertise/moonshot.md
How Agent Expertise Works
Each agent has a persistent expertise file that evolves:
Initial State
ceo.md
โโโ Role Description โ What this agent brings
โโโ Core Perspectives โ Their lens on problems
โโโ Decision History โ Table of all decisions they've participated in
โโโ Learning Notes โ What they've learned over time
After Each Deliberation
Agents update their files with:
Decision Record โ Topic, position taken, outcome
Learning โ What they learned from the debate
Pattern Recognition โ How this fits with past decisions
Accuracy Tracking โ How did their prediction fare?
Example Update
| Date | Topic | Position | Outcome | Notes |
|------|-------|----------|---------|-------|
| 24/03/2026 | Acquisition Offer | "Pursue, but negotiate harder" | TBD | Concerned about valuation, team retention |
Deliberation Quality Metrics
The board improves over time. Track:
Decision Quality โ How well do decisions age?
Agent Agreement โ Which agents tend to align?
Contrarian Accuracy โ How often does the Contrarian spot real risks?
Compounder Insight โ Which compound bets pay off?
Speed โ How quickly does the board reach consensus?
Best Practices
For Brief Writers
Be Specific โ Vague situations lead to weaker deliberations
Include Data โ Concrete metrics help agents reason better
State Constraints โ Be honest about limitations
Ask Real Questions โ Don't lead the board toward a predetermined answer
For Board Users
Trust the Process โ Diverse perspectives catch blindspots
Review Dissent โ The Contrarian's concerns often matter most
Track Predictions โ Come back and see how the board's predictions fared
Improve Expertise โ Manually update agent files with industry insights
For Maintenance
Keep Expertise Current โ Update agent files as patterns emerge
Archive Decisions โ Keep decision history for reference
Learn from Outcomes โ Track prediction accuracy
Customise the Oracle โ Keep custom-oracle.md current with Unite-Group's evolving context
The CEO Board integrates with the Unite-Group workflow:
Briefs can reference Linear issues or business objectives
Memos inform ROADMAP.md updates
Expertise files accumulate knowledge about the 7 businesses over time
Custom Oracle provides AUS regulatory and MACAS advisory context
Related Commands
/ceo-begin โ Start a board deliberation
/swarm-audit โ Audit codebase before committing to a technical decision
/hey-claude โ Load full project context before asking questions
Credits
Inspired by: IndyDevDan's agent scaling insights
Insight:"If you template your engineering, your agents can do exactly what you did. This is the big advantage. When you're not creating prescriptions, workflows, and systems for your agents to repeat, you miss out on all the true leverage."
