Library · topic
Orchestration
How agent work gets structured: single loops vs. supervisors, subagents, handoffs, and the observability to see what actually happened. The recurring lesson is restraint — coordination is overhead, and the best architecture is the simplest one that still meets the task. Start with the orchestration trade-offs note, then go deeper on the pattern you're actually considering.
- Deep diveverified 2026-07-28
Tracing and replay for agents
Represent each agent run as a trace with spans for model calls, retrieval, tool execution, guardrails, handoffs, and human approvals, linked by stable…
- Deep diveverified 2026-07-28
Multi-agent orchestration tradeoffs
Use multiple agents when subtasks are genuinely independent, require distinct tools or context, or benefit from parallel search or review; a single ag…
- Deep diveverified 2026-07-28
When should you use subagents?
A subagent is a specialized worker with its own isolated context window and restricted tools; the orchestrator hands off a bounded task and only the r…
- Field noteverified 2026-07-14
Production back-office agents
Back-office agents are useful when documents vary and judgment is needed, while deterministic RPA remains preferable for stable screens and fixed rule…