Agent Team

Agent Team

Also called: 智能体团队 · Agent squad

A set of agents in fixed, specialised roles collaborating persistently on a shared goal — a stable "team" rather than ad-hoc spawning.

Shared goalPlannerBuilderTesterCritic
Schematic (simplified)

Relation to multi-agent systems

"Agent team" is a productised framing of a multi-agent system: roles (planner, builder, tester, critic) are defined up front with their own prompts and tools, and the user faces a "team" rather than a set of loose processes.

Compared with ad-hoc subagents, the roles in a team are long-lived, reusable across tasks, and easier to give a fixed collaboration protocol (who acts first, who reviews whose output).

When it fits

When a task recurs with the same division of labour (write code → run tests → review → fix), freezing it into a team is more stable and predictable than having the lead agent improvise the structure each time, and it makes per-role tuning easier.

Common misconceptions

  • A "team" is an organising method, not a capability guarantee — blurry roles add coordination cost.
  • Team members do not automatically share context either; hand-offs must be defined explicitly.

Related terms

Sources

  1. Anthropic — Claude Docs: Claude Code subagents
  2. Wooldridge, M. — An Introduction to MultiAgent Systems (2009)

Compiled 2026-08-29 · This glossary is compiled from public papers, official specifications, and common industry definitions, and is updated as the field evolves. Corrections welcome.