multi-agent-experts

Use orchestration to spawn multiple specialized expert sessions via sessions_spawn, routing them with strict sessionKeys. This enables parallel evaluation and structured synthesis of analyzed outputs for comprehensive analyses.

Use multi-agent orchestration for complex, multi-dimensional questions requiring input from multiple expert perspectives. It applies when you need parallel evaluation and structured synthesis, such as strategic planning, risk analysis, and complex decision support.

Manage context by using explicit session-key based routing and reuse. The coordinator assigns unique sessionKeys to each spawned agent, maintaining context and feeding task descriptions to agents who return analyzed outputs.

Allocate tasks across specialized roles including researchers, thinkers, coaches, and other experts. The system handles intelligent expert matching and task allocation based on the analytical requirements of the complex question.

Multi-agent orchestration operates without external dependencies. It relies entirely on internal session-spawn orchestration logic and reusable task description protocols to dispatch agents and synthesize their returned analyzed outputs.

Multi-agent debates are limited to complex, multi-dimensional questions. Simple queries do not require orchestrated parallel evaluation or structured synthesis, making the session-spawn overhead unnecessary for straightforward analytical tasks.