vote-aggregator

# Vote Aggregator

Combine 5 independent judge verdicts into actionable final recommendation.

## Purpose

After all 5 judges (Pragmatist, Purist, Innovator, Skeptic, Optimizer) evaluate a proposal independently, this skill synthesizes their feedback into a clear go/no-go decision with actionable next steps.

## Input

```typescript
interface AggregatorInput {
  verdicts: JudgeVerdict[];  // Array of 5 verdicts
  parsedProposal: ParsedProposal;  // Original proposal for context
}

interface JudgeVerdict {
  judgeName: string;  // "Pragmatist", "Purist", etc.
  verdict: "APPROVE" | "REJECT";
  reasoning: string;
  keyConcerns: string[];
  suggestions: string[];
}
```

## Processing Steps

### Step 1: Count Votes

```markdown
Tally verdicts:
- Count APPROVE votes
- Count REJECT votes
- Track which judges voted which way

Example:
{
  approve: 3,
  reject: 2,
  judges: {
    approve: ["Pragmatist", "Innovator", "Optimizer"],
    reject: ["Purist", "Skeptic"]
  }
}
```

### Step 2: Determine Final Decision

```markdown
**Decision Matrix:**

5 APPROVE, 0 REJECT → APPROVED (unanimous)
4 APPROVE, 1 REJECT → APPROVED (strong consensus)
3 APPROVE, 2 REJECT → NEEDS_REVISION (split decision)
2 APPROVE, 3 REJECT → NEEDS_REVISION (major concerns)
0-1 APPROVE, 4-5 REJECT → REJECTED (strong opposition)

**Rationale:**
- 4-5 approvals = Clear green light, address concerns during implementation
- 3 approvals = Mixed signals, need to address rejecting judges' concerns first
- 0-2 approvals = Fundamental issues, requires redesign

Note: "NEEDS_REVISION" means "fix these specific things and resubmit"
```

### Step 3: Extract Common Themes

```markdown
Analyze all 5 verdicts to identify patterns:

**Strengths (Mentioned by 2+ judges):**
- Parse all `reasoning` and `suggestions` for positive points
- Look for common praise
- Example: Both Pragmatist and Innovator mention "good user value"

**Concerns (Mentioned by 2+ judges):**
- Parse all `keyConcerns` arrays
- Identify overlapping worr