multi_agent_collaboration

Multi-agent orchestration coordinates expert input from multiple agents by enabling staged collaboration and two-way chat when an agent's expertise is needed. It first lists available agents, then initiates a chat to involve them.

Task routing applies explicit decision rules for selecting agents and initiating chats. It lists available agents first, then uses two-way chat to query specialized capabilities or obtain workspace context, preserving session context via --session-id.

Yes, workspace context sharing is supported during cross-agent collaboration. You can obtain workspace or context from another agent by initiating a two-way chat after explicitly listing and selecting the right agent for the task.

To validate results with an expert agent, list available agents first, then start a two-way chat to involve them. This structured workflow ensures specialized validation while preserving session context via --session-id and avoiding callback to the source agent.

Yes, the orchestration workflow explicitly avoids callback to the source agent. It manages agent conversations with session tracking using --session-id, ensuring one-way consultation while preserving the full session context of the interaction.

Staged collaboration is needed when a user requests to involve an additional agent or when an agent's expertise is required for specialized work. It applies to scenarios requiring consulting another agent, obtaining context, or validating results.