//
Search across all documentation pages
11 pages in this section.
Learn to distinguish between single-agent loops and multi-agent systems. Understand when to use each architecture for optimal performance and cost.
Learn agentic orchestration basics with 10 Python examples, from minimal orchestrators to intermediate patterns, using the Claude Agent SDK.
Implement an orchestrator that dispatches work to Claude Agent SDK subagents and merges their results into a coherent output.
Learn to build subagents for parallel research by decomposing tasks, fanning out with asyncio.gather, and synthesizing findings into a coherent report.
Implement robust error recovery in agent loops. Learn to detect and respond to tool failures, stalled subagents, and bad results with retries, fallbacks, and circuit breakers.
Evaluate agent quality with this checklist before production. Learn to score task success, identify failure modes, and manage costs in multi-agent systems.
Design reliable, cost-bounded multi-agent architectures. Learn when to use multiple agents, design orchestrator/worker splits, and delegate work.
A single-page roundup of every highlight bullet from the 10 pages in the Agentic Orchestration section, grouped by source page so you can scan all 57 takeaways without opening each article individually.