agent-worker

Multi-agent orchestration coordinates AI agents across different backends by applying shared context, channel-based routing, and collaborative decision processes to manage scalable workflows. You can automate lifecycles using YAML workflows and CLI commands.

Channel-based routing supports @mention routing within multi-agent workflows, enabling targeted task delegation. Agents share channels and documents to facilitate collaborative tasks and voting within a unified programmatic environment.

Multi-agent orchestration supports SDK, Claude CLI, Codex, and Cursor backends. You can create, run, and manage single or multiple agents across these environments using command-line interfaces and YAML workflows.

Starting a multi-agent session with shared context involves using CLI commands to instantiate agents within a default workflow. You can create a new reviewer agent or manage shared documents for collaborative tasks programmatically.

Agent orchestration applies to both single-agent and multi-agent workflows, enabling programmatic control across multiple backends. It manages agent lifecycles and interactions through command-line interfaces and YAML workflows.

Automating AI agent lifecycles programmatically is best achieved through command-line interfaces and YAML workflows. This approach enables scalable, collaborative workflows and tool integration across multiple backends with shared context.