LLM Council Skill
A skill that asks ChatGPT and Gemini before Claude shows its plan, weighs the three views and delivers one plan crediting each model's contribution
- 482 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The script sends your question and task description to external OpenAI and Google services
- The OPENAI_API_KEY and GEMINI_API_KEY keys sit in plain text in the `.env` file and spend your money
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/llm-council-skillDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The script sends your question and task description to external OpenAI and Google services
- The OPENAI_API_KEY and GEMINI_API_KEY keys sit in plain text in the `.env` file and spend your money
This is third-party code. Review the repository files before installing.
What it does
LLM Council Skill lets Claude consult other models before it proposes a plan. You ask it to consult the council, ask ChatGPT and Gemini, or get perspectives from other AIs, and the skill runs the query_llms.py script. The script sends your question to ChatGPT and Gemini, Claude reviews the answers, picks out useful ideas, alternatives and concerns, adds its own analysis and assembles a single plan. At the end it briefly lists what each of the three models contributed. The script first tries the codex and gemini console clients installed on the system, because that saves tokens. If a client is missing, it calls the API using the OPENAI_API_KEY and GEMINI_API_KEY keys from a `.env` file or environment variables. The defaults are the cheap gpt-5-nano and gemini-3-flash-preview models, and stronger ones can be set in `.env`: gpt-5-mini, gpt-5.2, gemini-2.5-flash or gemini-3-pro-preview. If one service is unavailable, the skill says so and continues with the answers it received. Each API call has a 30-second timeout.
Who it is for. For developers and architects who want to cross-check a hard technical decision with several models before implementing it.
Good fit when
- You must choose a microservices architecture and want three positions instead of one
- You are designing a database schema and ask ChatGPT and Gemini to find weak spots
- A key decision is costly and a couple of calls for a second and third opinion are worth it
- You want to see where the models disagree: the final plan names each one's contribution
- You need a low-cost check: the default models are chosen for a low price
Not a fit when
- You need a fast answer to a simple question: each consultation makes two calls to external models
- Code and task descriptions must not go to third parties: the question goes to OpenAI and Google services
- You have neither API keys nor the codex and gemini clients installed: the script cannot reach the models
- You need a council of roles with debate and voting: this skill only gathers three opinions and runs no debate
Example request
Consult the council: How should I structure my React app for scalability?Limitations
The skill sends your question together with the task description to external OpenAI and Google services, so do not use it with code and data you must not disclose. The OpenAI API is unavailable from Russia, and keys for both models are paid for separately. Each council request makes two calls, the cost equals the sum of the chosen models' prices, and the author advises setting spending limits in the OpenAI and Google AI Studio dashboards. The `.env` file with keys must sit in the working folder. If the codex or gemini clients exist on the system, the script calls them first, and the cost then depends on your subscription. Installation is described as uploading the llm-council.skill file to Claude.
How to disable. Delete the llm-council skill in the Claude settings or its folder. The `.env` file with keys sits in your working folder, so delete it or revoke the keys in the OpenAI and Google AI Studio dashboards.
FAQ
How do I install LLM Council Skill?
Upload the llm-council.skill file from the repository to Claude, then copy `.env.template` to `.env` in the working folder and fill in the OpenAI and Gemini keys. The keys are needed only if the machine lacks the codex and gemini clients, which the script tries first.
Which phrases trigger a consultation?
The author's examples include "Consult the council", "Ask ChatGPT and Gemini", "Get perspectives from other AI models" and "Consult with other LLMs". The skill description asks it to fire when the user explicitly asks to consult other models before a plan.
Which models are used by default?
ChatGPT answers through gpt-5-nano and Gemini through gemini-3-flash-preview. Stronger options are set in `.env` with the OPENAI_MODEL and GEMINI_MODEL variables. For hard decisions the author suggests pairs such as gpt-5.2 and gemini-3-pro-preview.
What happens if one service does not respond?
The skill notes that this model's view is unavailable and continues with the others. The API timeout is 30 seconds.
How much does it cost?
Each council request makes two calls, so the price equals the sum of the two models' prices. The cheapest setups use gpt-5-nano and gemini-2.5-flash. The author advises setting monthly budgets in the OpenAI and Google AI Studio dashboards.
Discussion
No questions or reviews yet.
Related
SkillsA CLI and Python library for OpenAI, Claude, Gemini, local and other models: prompts, chats, tools, embeddings and a SQLite log
A Claude Code plugin that asks several AI models the same question and shows their answers side by side with a synthesis of agreement and divergence
Prompt caching audit
LLM Cache Audit Skill
Skill for Codex and Claude Code that traces what breaks LLM prompt caching in code and config and proposes a verification step
A skill that convenes a council of thinker personas and runs a structured deliberation on a hard decision, ending with a verdict
A skill that gives text-only models sight: an image by path or from a paste becomes structured JSON with transcribed text, layout and semantics
A Go command-line client for working with OpenAI, Azure, Perplexity and LLaMA language models right from the terminal
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Repository stars | Main difference |
|---|---|---|---|
| LLM Council Skill (this entry) | no | 482 | |
| claude-council | no | 911 | a Claude Code plugin that puts one question to several models and shows the answers beside an analysis |
| LLM Council | no | 2,825 | five advisors with different views and anonymous peer review, while here the outside voices are ChatGPT and Gemini |
README in short
The README describes what the skill does, shows dialogues in screenshots and lists trigger phrases. It then covers installation from the llm-council.skill file and key setup, tables of available OpenAI and Gemini models and recommended pairs from budget to professional. It ends with a cost estimate, savings tips, the skill's folder structure and a troubleshooting section: missing key, timeout, one API failing.
README badge
Original SKILL.md
--- 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**