Your AI-powered strategic advisor. Stop building. Start validating.

Install · How It Works · Examples · Components · Contributing

---

What is this?

CEO Review is an open-source Claude Code plugin that turns Claude into a tough, experienced CEO advisor who stress-tests your feature ideas before you write a single line of code.

Run /ceo-review and the plugin will:

1. Interrogate your idea through Socratic questioning — one tough question at a time
2. Dispatch 6 specialist analysts in parallel to evaluate strategy, market, moat, distribution, unit economics, and risk
3. Cross-check findings across analysts and re-run any that need updating based on new context
4. Deliver a scorecard with a clear BUILD / DEFER / KILL verdict

The goal is simple: catch bad ideas before they cost you weeks of engineering time.

Why?

Every team has shipped a feature that should have been killed in the planning phase. The post-mortem always reveals the same thing: nobody asked the hard questions early enough.

CEO Review forces those hard questions upfront. It applies real strategy frameworks — Hamilton Helmer's 7 Powers, Porter's Five Forces, TAM/SAM/SOM, Blue Ocean, JTBD — and demands specific metrics, not hand-waving. The tone is assertive and critical, but fundamentally wants you to succeed. Think of it as a co-founder who has been through this before and won't let you waste time on the wrong thing.

Installation

One-command install:

claude install-plugin https://github.com/tom-barkan/CEO-Advisor-Plugin

Manual install:

# Clone the repo
git clone https://github.com/tom-barkan/CEO-Advisor-Plugin.git

# Run Claude with the plugin
claude --plugin-dir /path/to/CEO-Advisor-Plugin

How It Works

Phase 1: Input

Provide a feature description as text or point to a spec file:

/ceo-review "Add AI-powered search to our documentation platform"
/ceo-review ./specs/ai-search-feature.md

Phase 2: Socratic Questioning

The CEO advisor asks 8-12 tough questions, one at a time. It adapts follow-up questions based on your answers and pushes back on vague responses.

Category	Example Question
Value	"How are users solving this problem TODAY without your feature?"
Strategy	"What are you saying NO to by saying yes to this?"
Financial	"How will you price this? Which metric does it move, and by how much?"
Distribution	"How does a user DISCOVER this feature?"
Risk	"Assume it launched and failed — why did it fail?"
Success	"What's the kill criteria — at what point do you pull the plug?"

If you give a vague answer, it pushes back once. If still vague, it flags it as a risk.

Phase 3: Parallel Analysis

Six specialist agents run simultaneously, each evaluating a different dimension:

Analyst	Focus	Key Output
Strategy Analyst	7 Powers, Porter's, Blue Ocean, Wardley	Strategic Alignment (1-10)
Market Scanner	Competition research, alternatives, saturation	Market Opportunity (1-10)
Moat Evaluator	Network effects, switching costs, defensibility	Defensibility (1-10)
Distribution Analyst	GTM channels, CAC, virality, adoption friction	Distribution Feasibility (1-10)
Unit Economics Analyst	TAM/SAM/SOM, pricing, LTV/CAC, metrics	Financial Value (1-10)
Pre-Mortem Analyst	Failure modes, risks, biggest assumptions	Risk Profile (1-10)

The Market Scanner uses web search to find real competitors — it names names, not hypotheticals.

Phase 4: Scorecard & Verdict

An orchestrator reviews all analyst findings, checks for contradictions, re-runs agents if new context warrants it, and produces the final scorecard:

SCORECARD
─────────────────────────────────
Strategic Alignment:       7/10  — Strong 7 Powers fit (network effects)
Market Opportunity:        5/10  — Crowded space, 4 funded competitors
Defensibility / Moat:      6/10  — Data moat possible but unproven
Distribution Feasibility:  8/10  — Strong PLG channel via existing users
Unit Economics:            4/10  — No pricing model defined
Risk Profile:              5/10  — High dependency on third-party API

VERDICT: DEFER
─────────────────────────────────
Strong distribution story but weak financial case.
Define pricing model and validate willingness-to-pay
before committing engineering resources.

Conditions for re-evaluation:
- Pricing survey with n=50+ target users
- Competitive differentiation beyond "we have it built in"
- Fallback plan for API dependency

Examples

# Review a feature idea from scratch
/ceo-review "Build a real-time collaboration feature for our design tool"

# Review from a spec file
/ceo-review ./specs/collaboration-feature.md

# Quick review with no arguments (will ask you to describe it)
/ceo-review

Components

Command

Command	Description
/ceo-review	Run a full CEO-level strategic review of a feature

Agents

Agent	Role	Color
review-orchestrator	Coordinates all analysts, synthesizes scorecard	Cyan
strategy-analyst	Strategic frameworks analysis (7 Powers, Porter's)	Blue
market-scanner	Competition research with web search	Cyan
moat-evaluator	Defensibility and competitive moat assessment	Green
distribution-analyst	Go-to-market and distribution feasibility	Yellow
unit-economics-analyst	Financial viability, TAM/SAM/SOM, pricing	Magenta
pre-mortem-analyst	Failure prediction and risk mitigation	Red

Skills

Skill	Purpose
strategic-frameworks	Reference knowledge for 7 Powers, Porter's, Blue Ocean, JTBD, Wardley Mapping, TAM/SAM/SOM, RICE/ICE, Kano Model
ceo-mindset	CEO persona definition, evaluation criteria, scorecard template, Socratic questioning patterns

Requirements

• Claude Code CLI installed
• Internet access (for market research via web search)

Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b my-feature)
3. Commit your changes
4. Open a pull request

Ideas for improvement:

• Add industry-specific scoring rubrics (SaaS, marketplace, consumer, enterprise)
• Integrate with project management tools to pull existing specs
• Add a "quick mode" that skips Socratic questioning for rapid scoring
• Historical tracking of past reviews and their outcomes
• Custom framework plugins (bring your own strategy framework)

License

MIT