# skill.cc/council

> **Summon domain experts for strategic evaluation.**

---

## Quick Start

```
Agent: Mount skill.cc/council, consult on [decision]
Human: "What would Feynman say?" / "Convene the council"
```

---

## What You Get

- **8 voice-imprinted advisors** — Each with distinct lens
- **Parallel evaluation** — Multiple perspectives on one question
- **Exportable deliberations** — Capture compute for memory
- **Customizable roster** — Add your own advisors

---

## Default Council

| ID | Name | Emoji | Lens |
|----|------|-------|------|
| `steve-crane` | Steve Crane | 🦅 | Cofounder practicality |
| `feynman` | Feynman | 🎯 | Bullshit detection, clarity |
| `paul-graham` | Paul Graham | 📝 | Startup patterns |
| `munger` | Charlie Munger | 📚 | Lollapalooza effects |
| `napoleon` | Napoleon | ⚔️ | Strategic territory |
| `pablos` | Pablos Holman | 🔧 | Deep tech energy |
| `vannevar` | Vannevar Bush | 📜 | Historical lineage |
| `danielle` | Danielle | 🌟 | Physics intuition |

---

## Usage

### Full Council
```
council.convene(question) → all 8 respond
```

### Individual
```
council.ask("feynman", question) → one response
council.ask(["feynman", "munger"], question) → subset
```

### With Context
```
council.convene(question, context_file)
```

### Export
```
council.convene(question).export() → markdown for memory
```

---

## Platform Routing

| Platform | Implementation |
|----------|----------------|
| Claude.ai | React artifact with proxied API calls |
| Claude Code | Sequential CLI prompts, markdown output |
| API | Structured JSON with all responses |
| MCP | Tool: `council_consult(question, members?)` |

---

## Customization

### Add Custom Member
```yaml
# Via skill.cc/craft
id: custom-advisor
name: "Domain Expert"  
emoji: "🔬"
voice: |
  You are [description].
  You evaluate through [lens].
  Your style is [style].
```

### Situation-Specific Councils
```yaml
# technical-review preset
members: [feynman, pablos, danielle]
context: "Technical architecture review"
```

---

## Philosophy

The council is not about getting "the right answer." It's about:

1. **Stress-testing** — Multiple lenses on one problem
2. **Externalizing** — Your thinking as distinct voices
3. **Capturing compute** — Deliberations become memory
4. **Pattern matching** — What would [expert] actually say?

---

## Related Skills

- `skill.cc/memory` — Store council deliberations
- `skill.cc/persist` — Auto-save to Mind Palace
- `skill.cc/think` — Extended reasoning

---

*Parent: skill.cc*