debate-templates

A multi-model adversarial debate workspace is a structured directory layout that organizes context, per-model round files, syntheses, and consensus artifacts to coordinate several AI agents debating a core decision.

Set up a multi-model debate by creating a context.md file, initializing per-model round-1 files in the predefined workspace layout, and seeding the first round prompts to start the adversarial evaluation flow.

Yes, the templates support round-based AI agent simulations by providing role-aware prompts and structured workflows for Round 1 and subsequent rounds, enabling clear evaluation, repeatability, and synthesis of debate outcomes.

The best way to organize multi-model debate artifacts is using the enforced fixed directory layout and naming conventions, which separate context, per-model responses, syntheses, and consensus artifacts for reproducible experiments.

No specific frameworks are required as dependencies; the templates provide the orchestration flow and reusable reference materials needed to run multi-model adversarial debates directly within your workspace.