consensus
Read-only
Structured multi-model debate for complex decisions: consult models with for/against stances, capture findings, and synthesize a consensus recommendation.
Instructions
Builds multi-model consensus through systematic analysis and structured debate. Use for complex decisions, architectural choices, feature proposals, and technology evaluations. Consults multiple models with different stances to synthesize comprehensive recommendations.
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| step | Yes | Consensus prompt. Step 1: write the exact proposal/question every model will see (use 'Evaluate…', not meta commentary). Steps 2+: capture internal notes about the latest model response—these notes are NOT sent to other models. | |
| images | No | Optional absolute image paths or base64 references that add helpful visual context. | |
| models | No | User-specified roster of models to consult (provide at least two entries). User-specified list of models to consult (provide at least two entries). Each entry may include model, stance (for/against/neutral), and stance_prompt. Each (model, stance) pair must be unique, e.g. [{'model':'gpt5','stance':'for'}, {'model':'pro','stance':'against'}]. When the user names a model, you MUST use that exact value or report the provider error—never swap in another option. Use the `listmodels` tool for the full roster. Top models: gemini-2.5-pro (score 100, 1.0M ctx, thinking, code-gen); gemini-3-pro-preview (score 100, 1.0M ctx, thinking, code-gen); gemini-2.5-flash (score 61, 1.0M ctx, thinking); gemini-2.0-flash (score 56, 1.0M ctx, thinking); gemini-2.0-flash-lite (score 42, 1.0M ctx). | |
| findings | Yes | Step 1: your independent analysis for later synthesis (not shared with other models). Steps 2+: summarize the newest model response. | |
| step_number | Yes | Current step index (starts at 1). Step 1 is your analysis; steps 2+ handle each model response. | |
| total_steps | Yes | Total steps = number of models consulted plus the final synthesis step. | |
| relevant_files | No | Optional supporting files that help the consensus analysis. Must be absolute full, non-abbreviated paths. | |
| continuation_id | No | Unique thread continuation ID for multi-turn conversations. Works across different tools. ALWAYS reuse the last continuation_id you were given—this preserves full conversation context, files, and findings so the agent can resume seamlessly. | |
| model_responses | No | Internal log of responses gathered so far. | |
| next_step_required | Yes | True if more model consultations remain; set false when ready to synthesize. | |
| current_model_index | No | 0-based index of the next model to consult (managed internally). | |
| use_assistant_model | No | Use assistant model for expert analysis after workflow steps. False skips expert analysis, relies solely on your personal investigation. Defaults to True for comprehensive validation. |