install-llm-council

What This Skill Does

LLM Council is an advanced orchestration skill designed to transform a single query into a consensus-driven dialogue. Instead of relying on the perspective of a single model, this skill queries multiple LLMs simultaneously, facilitates an internal critique phase where models evaluate each other's reasoning, and produces a final, synthesized chairman answer. It is built to be OpenClaw-native, meaning it eliminates the tedious configuration process by automatically resolving credentials via your existing OpenClaw workspace environment, local gateway, or OpenRouter API keys without manual prompts.

Installation

To install this skill, use the command: clawhub install openclaw/skills/skills/jeadland/install-llm-council. Once installed, the primary interface is the slash command /install-llm-council. This command triggers an automated pipeline that clones the required repository, manages environment-specific dependencies (using uv for Python and npm for frontend components), and writes a hardened .env file to ensure the backend and frontend are ready for immediate use. You can customize the deployment using the --mode flag (choosing between auto, dev, or preview) or specify a target directory using --dir.

Use Cases

• Research and Synthesis: Use this when you need a balanced view on complex, subjective, or multi-faceted topics where a single model might have a bias or blind spot.
• Fact Checking: By having models critique each other, you can surface hallucinations or inconsistencies in reasoning that one model might have introduced.
• Decision Making: When weighing pros and cons for a project, the "chairman" output provides a clear, reconciled recommendation based on the combined intelligence of your chosen model lineup.

Example Prompts

1. "/install-llm-council --mode preview"
2. "/install-llm-council --dir ~/projects/research-council"
3. "/install-llm-council"

Tips & Limitations

For the best results, ensure your OpenClaw gateway is active if you are relying on it for API access. Because this tool orchestrates multiple API calls, monitor your token usage carefully, as the critique-cycle generates significant traffic compared to standard chat queries. If you encounter port conflicts, the installer is designed to attempt a fallback, but ensuring ports 8001, 5173, and 4173 are free will lead to a smoother setup process. Always check the printed access URLs after installation to navigate directly to the dashboard.