autoconference:plan

Configuring an autonomous multi-agent research conference requires an interactive 8-step wizard that collects research goals, evaluation modes, and researcher setups to safely generate a fully populated conference.md file.

A metric-mode evaluator dry-run in research planning validates your numerical evaluation setup before execution, ensuring your target accuracy or scoring metrics are correctly configured to prevent wasted compute.

Setting up peer review and synthesis rounds for autonomous research involves defining researcher roles, search-space partitioning, and an optional Devil's Advocate mandate through a structured configuration generation process.

Yes, you can use rubric-based qualitative evaluation instead of numerical metrics, allowing progress to be judged by structured peer review rather than measured by a metric evaluator during the autonomous research conference.

Preventing wasted compute in multi-agent research requires interactively capturing a specific, measurable research goal and validating runtime constraints to produce a safe, correct configuration file before execution.

For an autonomous research conference, you can set runtime constraints including pause cadence, time budget, round limits, and iteration limits to control the execution strategy and behavior of the multi-agent system.