Agent handoff

Agent handoff is the multi-agent pattern where one agent decides another specialised agent should take over the task — transferring the conversation state to the new agent's context.

Multi-agent systems in 2026 (OpenAI Swarm, CrewAI, LangGraph hierarchies, Claude sub-agents) use handoffs as their core dispatch primitive. A coordinator agent identifies that the task belongs to a specialist (the billing agent, the code agent, the research agent), summarises the context that matters, and hands off. The receiving agent picks up with focused tools and a focused system prompt. Handoffs avoid the problem of monolithic agents with too many tools mis-routing. The risk: handoff loops where agents bounce work back and forth. Mitigation: max-handoff caps, no-progress detection, and an explicit "answer the user now" terminal action.

When to use agent handoff

• Multi-agent systems with specialised sub-agents.
• Customer support agents that route to billing / technical / sales specialists.

Common mistakes

• Handoff loops — set max-handoff caps.
• Losing critical context across handoff — summarise explicitly.

FAQ

Related terms

• AI agent — An AI agent is a system where a language model autonomously plans and executes a sequence of tool calls to accomplish a goal.
• Tool router — A tool router is a layer in an agent that decides which tool to call (or which sub-agent to delegate to) for a given step — distinct from a model router which picks the underlying LLM.
• Agent loop — An agent loop is the repeating cycle of an AI agent — observe state, decide on an action (usually a tool call), execute, observe the result, and repeat — until a goal is reached or a stop condition fires.
• Stateful agent — A stateful agent persists state — memory, learned facts, long-running context — across sessions, in contrast to stateless agents that start fresh on every conversation.

Last updated: 2026-06-01. Raw markdown: https://promtable.com/glossary/agent-handoff.md.