Stage: CRITIQUE (Spec Critique)

Deep critical analysis of a draft or full spec to identify architectural flaws, hidden complexity, and implementation risks.

This is different from PRE-CHECK:

• PRE-CHECK: Can we build this at all? (APIs, legal, costs)
• CRITIQUE: Is this spec well-designed? (architecture, complexity, scope)

Purpose

Find issues that would only surface "6 feet into development":

• Architectural flaws
• Underestimated complexity
• Missing edge cases
• Unrealistic promises
• Scope creep indicators

Prerequisites

• For draft critique: outputs/{name}/draft.md
• For spec critique: outputs/{name}/spec.md

Parameters

• --name: Project name to critique
• --stage: draft or spec (default: auto-detect)

Process

For Draft Critique

Focus on high-level viability:

1. Problem-Solution Fit

□ Does the solution actually solve the stated problem?
□ Is the problem real or assumed?
□ Are there simpler solutions we're missing?

2. Feature Scope

□ Are features truly MVP or scope creep?
□ Can each feature be built in days, not weeks?
□ Are there hidden dependencies between features?

3. Technical Approach

□ Is the architecture appropriate for the problem?
□ Are there obvious alternatives not considered?
□ What's the complexity we're not seeing?

4. Monetization Reality

□ Will users pay this for this value?
□ Do the unit economics work?
□ Is the pricing competitive?

For Spec Critique (Full Analysis)

Everything from draft critique PLUS:

5. API/Integration Deep Dive

□ For each API call in spec:
  - What's the actual response format?
  - What errors can occur?
  - What are the edge cases?
□ Do the tools match what the API can do?
□ Are there API limitations not acknowledged?

6. Architecture Stress Test

□ What happens at 10x scale?
□ What happens with bad input?
□ What happens when dependencies fail?
□ State management: is "stateless" actually stateless?

7. Feature Feasibility

□ For each feature:
  - Is this actually possible with the stated approach?
  - What complexity is hidden in "just do X"?
  - What's the real implementation effort?
□ Are there features that contradict each other?

8. Promise vs Reality

□ What does the spec promise that we can't deliver?
□ What UX expectations are unrealistic?
□ Where are we hand-waving complexity?

Output

Saves to outputs/{name}/critique.md:

# Critique: {name}

## Critique Type: {draft | spec}
## Date: {date}

## Risk Summary

| Category | Risk Level | Key Concern |
|----------|------------|-------------|
| Problem-Solution Fit | 🟢/🟡/🔴 | ... |
| Feature Scope | 🟢/🟡/🔴 | ... |
| Technical Approach | 🟢/🟡/🔴 | ... |
| API Feasibility | 🟢/🟡/🔴 | ... |
| Hidden Complexity | 🟢/🟡/🔴 | ... |
| Monetization | 🟢/🟡/🔴 | ... |

## Decision

[ ] **PROCEED** - Spec is solid, move to next stage
[ ] **REVISE** - Issues found, update spec and re-critique
[ ] **PIVOT** - Fundamental issues, need different approach
[ ] **KILL** - Spec reveals this isn't viable

## Critical Issues (Must Address)

### Issue 1: {title}
- **What:** Description of the issue
- **Why it matters:** Impact if not addressed
- **Recommendation:** How to fix

### Issue 2: ...

## Warnings (Should Address)

### Warning 1: {title}
- **What:** Description
- **Risk level:** Low/Medium
- **Recommendation:** Suggested action

## Spec Corrections Needed

| Section | Current | Should Be |
|---------|---------|-----------|
| Feature X | "Auto-resolve linked records" | "Resolve linked records (1 level, 2-3 API calls)" |
| ... | ... | ... |

## If Pivoting

See: `/mcp-opportunity-pipeline:pivot --name {name}`

---

*Critique completed: {date}*

Decision Criteria

PROCEED

• No critical issues
• Warnings are acceptable risks
• Architecture is sound
• Scope is realistic

REVISE

• Critical issues found but fixable
• Spec needs corrections
• Re-run critique after updates

PIVOT

Use pivot skill to analyze options:

• Same problem, different technical approach
• Same idea, reduced scope
• Related idea, different angle

KILL

• Fundamental architectural flaw
• Scope is actually months, not weeks
• Monetization doesn't work after analysis

Usage

/mcp-opportunity-pipeline:critique --name notion-database-sync-mcp --stage draft
/mcp-opportunity-pipeline:critique --name airtable-advanced-mcp --stage spec

Integration with Pivot

When decision is PIVOT, use the pivot-analysis skill:

Use skill: pivot-analysis

Analyze pivot options for: {name}
Current stage: {draft | spec}
Reason for pivot: {from critique}

The skill will provide:

• Pivot type recommendations
• Effort comparison
• Risk assessment for each path
• Recommended next action

Next Steps

• If PROCEED (draft) → /mcp-opportunity-pipeline:spec
• If PROCEED (spec) → /mcp-opportunity-pipeline:walkthrough
• If REVISE → Update spec, re-run critique
• If PIVOT → /mcp-opportunity-pipeline:pivot
• If KILL → Document and move on