research

Multi-stage research workflows are automated by coordinating parallel scientist agents that decompose complex goals, execute independent investigations, verify findings, and synthesize comprehensive reports across codebases, data, and literature.

Parallel research agent orchestration is the process of decomposing complex goals into independent stages and running concurrent investigations. You need it for tasks requiring simultaneous execution, verification, and synthesis across diverse domains like codebases and literature.

You run autonomous research exploration by enabling AUTO mode, which allows parallel scientist agents to autonomously decompose goals, execute multi-stage investigations, verify findings, and synthesize results without manual intervention until the task is complete.

Yes, you can use parallel scientist agents to execute independent investigations and automatically synthesize verified findings into a structured report. The orchestration supports staged execution to ensure comprehensive synthesis across all research stages.

Yes, this research orchestration works simultaneously for analyzing codebases, data, and literature. It decomposes complex goals into independent stages, allowing parallel agents to investigate across these distinct domains before verifying and synthesizing the combined results.

The main limitation of using autonomous agents for research decomposition is the dependency on the AUTO mode's ability to accurately interpret complex goals. While it supports autonomous exploration, highly ambiguous objectives may require explicit frontmatter metadata and configurable agent prompts to guide the staged orchestration effectively.