Agentic Architecture & Orchestration

The core agentic loop pattern: send a request, inspect stop_reason, execute tools, return results, and repeat. Master the lifecycle that powers every Claude agent and learn exactly how stop_reason drives loop control -- the single most tested concept in Domain 1.

Hub-and-spoke architecture where a coordinator agent manages all inter-subagent communication, decomposes tasks, delegates to specialized subagents, and synthesizes results. Understand why subagents have isolated context and how the coordinator prevents coverage gaps.

How to spawn subagents using the Task tool, pass context explicitly (since nothing is inherited), configure AgentDefinitions with restricted tool sets, and leverage fork-based session management for parallel exploration.

Designing multi-step agent workflows with programmatic enforcement of prerequisites, structured handoff protocols for human escalation, and strategies for decomposing multi-concern customer requests into parallel investigations. Learn why prompt-based guidance alone is insufficient for critical compliance.

Using Agent SDK hooks to intercept and transform tool results, enforce business rules at the code level, and implement deterministic compliance guarantees. Learn PostToolUse hooks, tool call interception, and when to choose hooks over prompt instructions.

Breaking complex tasks into manageable subtasks using two fundamental strategies: fixed sequential pipelines (prompt chaining) for predictable workflows, and dynamic adaptive decomposition for open-ended tasks. Learn when to use each and how to avoid attention dilution.

Managing session state across long-running and multi-session investigations. Named session resumption with --resume, fork_session for parallel exploration branches, and strategies for deciding between resuming stale sessions and starting fresh with structured summaries.