research

You can automate multi-stage research workflows by orchestrating parallel scientist agents that decompose complex goals into concurrent stages, running independent investigations and verification loops across complex codebases. This enables faster insights through parallel execution.

AUTO mode is an autonomous control mechanism that runs research sessions end-to-end until completion without manual intervention. It manages stage decomposition, parallel invocation, and verification loops automatically to satisfy research goals.

Start a parallel research session by providing a goal, use AUTO mode for end-to-end autonomous runs, check progress with status commands, resume the last session if interrupted, and generate synthesized reports using the session ID.

Yes, you can run independent investigations concurrently by decomposing complex goals into parallel stages. The system applies verification loops to cross-validate findings across parallel scientist agents before generating synthesized, traceable outputs.

Yes, hypothesis testing and cross-validation are core use cases supported by parallel scientist agents. The system orchestrates independent investigations concurrently and applies verification loops to validate findings across complex codebases.

Use parallel research agents when working with multi-stage research workflows, hypothesis testing, or complex codebases where independent investigations can run concurrently. Sequential execution is less efficient when stages can be decomposed and executed in parallel.