configure_council
Update council configuration: select models, set a judge, choose response mode, and adjust deconflict rounds.
Instructions
Update the council configuration: select which models form the council, choose a judge model, set the response mode (individual / categorized / deconflicted), and set the maximum deconfliction rounds.
Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| models | No | Council member model IDs. Format: "provider:model" or "provider/serverId:model". Examples: "ollama:llama3", "openai:gpt-4o", "vllm/server1:meta-llama/Llama-3-8B" | |
| judge_model | No | Judge model ID. Same format. Omit for auto (largest council member). | |
| auto_council | No | Default true. When true and no models are set, auto-populate the council from all available Ollama chat models (local + :cloud). | |
| response_mode | No | individual: raw responses. categorized: agreement/complementary/conflicting. deconflicted: iterative loop with deconfliction score. | |
| max_deconflict_rounds | No | Max deconfliction rounds (1–10, default 3). |