research

You can automate parallel research tasks by orchestrating scientist agents with AUTO mode, which handles decomposition, concurrent execution, verification, and synthesis autonomously. Define your research goal explicitly and run it to reduce manual coordination.

AUTO mode enables fully autonomous research execution from goal decomposition to final synthesis. It manages loop control, state persistence, and progress reporting across parallel agents without requiring manual intervention at each stage.

You can check progress of an active parallel research session by running the status command. This reports on the current state of decomposed tasks, agent execution, and verification stages within your ongoing autonomous session.

Yes, the system breaks complex research goals into independent stages and runs them concurrently across agents. It cross-validates findings through verification and synthesizes a comprehensive comparative report with structured outputs.

No, explicit goal decomposition is handled automatically by the orchestration layer. You provide a structured task definition and model prompts, and the system partitions the goal into independent stages for concurrent agent execution.

Autonomous orchestration requires explicit task definitions, model prompts, and structured session management to function correctly. Without clear goal specifications, the decomposition and parallel execution stages may not produce coherent synthesis results.