# llm-council

Adopt this skill: https://proskills.md/skills/other/gcpdev-llm-council-skill-llm-council

Page: https://proskills.md/skills/other/gcpdev-llm-council-skill-llm-council
This document is the public skill bundle for that page. It contains SKILL.md and every helper file the catalog carries. No login is required.

## File: llm-council/SKILL.md
size: 4184
upstream: https://github.com/gcpdev/llm-council-skill/blob/main/llm-council/SKILL.md

````text
---
name: llm-council
description: Multi-LLM collaborative brainstorming and planning. Use when user explicitly requests consultation with multiple AI models (ChatGPT, Gemini, other LLMs) before presenting an implementation plan, or asks to "consult the council", "ask other models", or "get perspectives from other AIs". Queries external LLM APIs, synthesizes their perspectives, and presents an adapted implementation plan.
---

# LLM Council

Consult multiple AI models (ChatGPT and Gemini) for their perspectives before presenting implementation plans to users.

## Workflow

When user requests consultation with other AI models, use phrases like:
- "Consult with ChatGPT and Gemini about..."
- "Ask other AI models what they think about..."
- "Get perspectives from the council on..."
- "Consult the LLM council: [your question]"

**Process:**

1. **Query external LLMs**: Run `scripts/query_llms.py` with the user's prompt to get perspectives from both ChatGPT and Gemini
2. **Analyze responses**: Review what each model suggests, identifying valuable insights, alternative approaches, and potential concerns
3. **Synthesize plan**: Create an implementation plan that incorporates the best ideas from all three models (Claude's own analysis + ChatGPT + Gemini)
4. **Present to user**: Show the final plan along with a brief summary of key contributions from each model

## Setup Requirements

The skill requires API keys and optional model configuration stored in a `.env` file in the working directory:

```
OPENAI_API_KEY=sk-...
GEMINI_API_KEY=...

# Optional: Specify which models to use (defaults shown below)
OPENAI_MODEL=gpt-5-nano
GEMINI_MODEL=gemini-3-flash-preview
```

**Default Models:**
- ChatGPT: `gpt-5-nano` (fastest, most cost-efficient - $0.05/1M input, $0.40/1M output)
- Gemini: `gemini-3-flash-preview` (balanced speed and intelligence)

**Upgrade Options for Better Collaboration:**

*OpenAI models (ordered by capability and cost):*
- `gpt-5-nano` - Fastest, most cost-efficient ($0.05/1M in, $0.40/1M out) - **DEFAULT**
- `gpt-5-mini` - Faster, cost-efficient for well-defined tasks ($0.25/1M in, $2.00/1M out)
- `gpt-5.2` - Best for coding and agentic tasks ($1.75/1M in, $14.00/1M out)
- `gpt-5.2-pro` - Smarter, more precise for complex problems ($21.00/1M in, $168.00/1M out)

All models support reasoning tokens, 400K context window, and image input.

*Gemini models (ordered by capability):*
- `gemini-2.5-flash-lite` - Ultra-fast, optimized for throughput
- `gemini-2.5-flash` - Best price-performance, large-scale processing
- `gemini-3-flash-preview` - Balanced speed and frontier intelligence (default)
- `gemini-3-pro-preview` - Most intelligent multimodal model, best for complex reasoning

Higher-tier models provide more sophisticated analysis but cost more per API call.

If the `.env` file doesn't exist or keys are missing, inform the user and provide setup instructions.

## Usage Example

**User input:** "Consult the council: How should I architect a real-time data pipeline for IoT sensors?"

**Claude's process:**
1. Execute: `python3 scripts/query_llms.py "How should I architect a real-time data pipeline for IoT sensors?"`
2. Parse JSON responses from ChatGPT and Gemini
3. Analyze their suggestions (e.g., ChatGPT suggests Kafka, Gemini recommends considering edge computing)
4. Synthesize final plan incorporating valuable insights from all models
5. Present the adapted plan to user with attribution

## Output Format

Present the final implementation plan naturally, mentioning key insights from other models inline where relevant. For example:

"Based on consultation with ChatGPT and Gemini, here's the recommended architecture:

[Implementation plan with inline references like "ChatGPT highlighted the importance of..." or "Gemini suggested..."]

Key contributions:
- ChatGPT: [brief summary]
- Gemini: [brief summary]"

## Error Handling

- If API keys are missing, inform user and provide setup instructions
- If an API call fails, note which model's perspective is unavailable and proceed with available responses
- If both APIs fail, inform user and offer to provide Claude's own analysis without external consultation
````

## File: README.md
upstream: https://github.com/gcpdev/llm-council-skill/blob/main/README.md

````text
# LLM Council Skill

A Claude skill that enables collaborative brainstorming with multiple AI models (ChatGPT and Gemini) before presenting implementation plans.

## What It Does

When you ask Claude to consult with other AI models, Claude will:
1. Query both ChatGPT and Gemini for their perspectives
   <img width="607" height="210" alt="council_demo" src="https://github.com/user-attachments/assets/6e2a759d-40cc-4bc2-9651-a74aca651303" />
3. Analyze their responses and identify valuable insights
4. Synthesize a comprehensive implementation plan incorporating ideas from all three models
5. Present the final plan with attribution to each model's contributions
   <img width="608" height="222" alt="council_demo2" src="https://github.com/user-attachments/assets/327893bd-9806-4cfb-85df-538649c978fa" />

## How to Use

Simply ask Claude to consult with other AI models using phrases like:

- "**Consult the council:** How should I architect a microservices system?"
- "**Ask ChatGPT and Gemini** what they think about my database design"
- "**Get perspectives from other AI models** on this technical decision"
- "**Consult with other LLMs:** What's the best approach for..."
   <img width="609" height="79" alt="council_demo3" src="https://github.com/user-attachments/assets/ba98d179-08be-4cc8-ac6a-0e1bc69b3fc7" />

**Example:**
```
User: Consult the council: How should I structure my React app for scalability?

Claude will then:
- Query ChatGPT and Gemini about React architecture
- Analyze their suggestions on components, state management, and organization
- Present a synthesized plan incorporating insights from all three models
```

## Installation

1. **Install the skill** in Claude by uploading the `llm-council.skill` file
2. **Set up API keys and model preferences**:
   - Copy `.env.template` to create a `.env` file in your working directory
   - Add your OpenAI API key (get it at https://platform.openai.com/api-keys)
   - Add your Gemini API key (get it at https://aistudio.google.com/app/apikey)
   - Optionally customize which models to use (see Model Options below)

## Model Options

**Default Models (Fast & Cost-Effective):**
- ChatGPT: `gpt-5-nano-2025-08-07` (highly cost-effective)
- Gemini: `gemini-3-flash-preview` (balanced speed and intelligence)

**Upgrade Options for Better Collaboration:**

**OpenAI models (ordered by capability):**
- `gpt-5-nano` - Fastest, cheapest version of GPT-5. It's great for summarization and classification tasks. (Default)
- `gpt-5-mini` - Balanced cost and quality
- `gpt-5.2` - Smart model, capable of most tasks
- `gpt-5.2-pro` - State-of-the-art for professional knowledge work

**Gemini models (ordered by capability):**
- `gemini-2.5-flash-lite` - Ultra-fast, optimized for throughput
- `gemini-2.5-flash` - Best price-performance
- `gemini-3-flash-preview` - Balanced (default)
- `gemini-3-pro-preview` - Most intelligent, best reasoning

**How to Configure:**
Add these lines to your `.env` file:
```
OPENAI_MODEL=gpt-5-nano
GEMINI_MODEL=gemini-3-flash-preview
```

**Recommended Configurations:**
- **Balanced**: Defaults (`gpt-5-nano` + `gemini-3-flash-preview`)
- **Budget**: `gpt-5-nano` + `gemini-2.5-flash`
- **High Quality**: `gpt-5` + `gemini-3-flash-preview`
- **Premium Reasoning**: `gpt-5.2` + `gemini-3-pro-preview`
- **Professional Work**: `gpt-5.2-pro` + `gemini-3-pro-preview`

## Benefits

- **Diverse perspectives**: Get insights from three different AI models with different training and capabilities
- **Better decisions**: Identify potential issues or alternatives you might have missed with a single model
- **Comprehensive planning**: Combine strengths of multiple models for more robust implementation plans

## API Costs

Both OpenAI and Gemini APIs have usage costs that vary significantly by model:

**OpenAI Cost Tiers (approximate, check current pricing):**

*Budget Tier:*
- `gpt-5-nano`: Very low cost per token
- 
*Standard Tier:*
- `gpt-5-mini`: Moderate cost

*Premium Tier:*
- `gpt-5.2`: Higher cost per token
- `gpt-5.2-pro`: Highest cost for professional work

**Gemini Cost Tiers (approximate, check current pricing):**

*Budget Tier:*
- `gemini-2.5-flash-lite`: Very low cost, optimized for throughput
- `gemini-2.5-flash`: Low cost, best price-performance

*Standard Tier:*
- `gemini-3-flash-preview`: Moderate cost (default)

*Premium Tier:*
- `gemini-3-pro-preview`: Higher cost for advanced reasoning

**Cost Management Tips:**
- Start with default models for routine brainstorming (very cost-effective)
- Use mid-tier models (`gpt-5-mini` + `gemini-3-flash-preview`) for balanced quality/cost
- Upgrade to premium models only for critical architectural decisions or complex reasoning tasks
- Set usage limits in your API dashboards (OpenAI Platform and Google AI Studio)
- Consider setting monthly budgets to avoid surprises
- Monitor your usage patterns and adjust model choices accordingly

**Note:** Each `/council` command makes 2 API calls (one to ChatGPT, one to Gemini), so total cost is the sum of both models' pricing.

## Skill Structure

```
llm-council/
├── SKILL.md                    # Main skill instructions
├── scripts/
│   └── query_llms.py          # Python script that queries both APIs
└── references/
    └── setup.md               # Detailed setup instructions
```

## Troubleshooting

**"API key not found" error**: Make sure your `.env` file is in the current working directory with the correct keys.

**API timeout**: The script has a 30-second timeout per API. If an API is slow or down, it will show an error but continue with the other model's response.

**One API fails**: Claude will note which model's perspective is unavailable and proceed with the available responses.
````
