Teams of agents

An agent can hand work to other agents. The one that delegates is a manager; the ones it calls are its sub-agents. Together they are a team.

Not the same as a team of people

An organization's teams are groups of people you share things with. This page is about agents calling agents.

When it helps

Delegation is worth it when the work splits into jobs that need different instructions, tools or models: a manager that plans and writes, a researcher that reads the web, a specialist that knows one system well. It is not worth it for "do the same thing twice" — that costs two model calls for one result.

Setting it up

In Configure, add sub-agents to the manager. Then:

• The manager's model must support tools: delegation is a tool call.
• Sub-agents must be published, and belong to the same owner. Work never crosses from one owner to another.
• A chain can be three deep, and cycles are refused: an agent cannot end up calling itself.
• A sub-agent follows its latest published version, so improving it improves every manager that uses it.

What the person sees

In chat, a delegation appears as a card: which agent was asked, what it was asked, and what it answered. Nested delegation is indented. Nothing happens off-screen.

What it costs

A delegated answer is a full run of the other agent: its own model calls, its own tools, its own tokens. Budgets apply at each step, and the whole tree shares one limit on steps and tokens, so a confused team stops rather than spiralling.

Watch the trace the first few times. The usual surprise is a manager that delegates far more often than you expected — the fix is in the manager's instructions, saying plainly when to do the work itself.