Skip to content
DocsUse AIworkflowsTeam Orchestration
Chapter 10 of 15·workflows·11 min read

Team Orchestration

Điều Phối Team

Multi-agent coordination for complex projects

Hover or tap any paragraph to see Vietnamese translation

Multi-Agent Orchestration

For complex projects, a single agent may not be enough. Claude Code supports orchestrating multiple agents working in parallel, each specializing in a different area.

Team orchestration allows you to divide work into independent tasks, assign them to appropriate agents, and combine results into a complete product.

Team Orchestration Workflow
1
Create Team
/team create
2
Define Tasks
Break down work
3
Assign Agents
Match expertise
4
Execute
Parallel work
5
Combine
Merge results

Specialized Agents

Architect
System design
Requirements analysis
Planning
Executor
Write code
Create files
Run commands
Code Reviewer
Find bugs
Check best practices
Performance
Security Reviewer
OWASP Top 10
Secret detection
Unsafe patterns

Architect Agent

Specializes in system architecture design, requirements analysis, and implementation planning. Uses Opus model for deep analysis.

Executor Agent

Performs specific tasks like writing code, creating files, running commands. Can use Haiku (simple), Sonnet (medium), or Opus (complex).

Code Reviewer Agent

Reviews code to find bugs, best practice violations, and performance issues. Provides structured feedback with severity ratings.

Security Reviewer Agent

Specializes in finding security vulnerabilities per OWASP Top 10, exposed secrets, and unsafe patterns.

Using Team Command

Team Command
1# Create a team for a project2/team create "feature-auth"34# Team will:5# 1. Create shared task list6# 2. Spawn specialized agents7# 3. Coordinate work in parallel8# 4. Report back when complete910# Example team workflow:11Leader: Creates tasks, assigns to agents12Agent 1 (Executor): Implements backend API13Agent 2 (Executor): Implements frontend UI14Agent 3 (Reviewer): Reviews completed work

Task List Coordination

Agents in a team share a task list. Each agent can claim tasks, mark them complete, and create new tasks when additional work is discovered.

Task List
1# Task list example2#1. [in_progress] Create user schema (owner: agent-1)3#2. [pending] Add login endpoint4#3. [pending] Add signup endpoint5#4. [blocked] Add OAuth - blocked by #2, #36#5. [completed] Setup database migrations

Parallel Execution

When tasks don't depend on each other, they can be executed in parallel by different agents, significantly speeding up completion.

Parallel Execution
1# Independent tasks run in parallel:2┌──────────────────────────────────────────┐3│ Time  │ Agent 1      │ Agent 2           │4├───────┼──────────────┼───────────────────┤5│ 0:00  │ User API     │ Product API       │6│ 0:05  │ User tests   │ Product tests     │7│ 0:08  │ (idle)       │ (idle)            │8└──────────────────────────────────────────┘910# Sequential: 0:1611# Parallel:   0:08 (2x faster)

Communication Between Agents

Agents communicate through the SendMessage tool. Leader agent coordinates and agents report progress.

Agent Communication
1# Agent communication2Leader → Agent-1: "Implement user registration API"3Agent-1 → Leader: "Task complete, created POST /users"4Leader → Agent-2: "User API ready, you can start frontend"
Tip
The leader agent should check progress regularly and adjust plans if needed.

When to Use Teams

  • Large projects with multiple independent modules.
  • Need different expertise (code, review, security).
  • Want to speed up through parallel execution.

Key Takeaways

Điểm Chính

  • Teams coordinate multiple specialized agents working in parallelTeams điều phối nhiều agent chuyên biệt làm việc song song
  • Shared task list keeps everyone synchronizedTask list chung giữ mọi người đồng bộ
  • Independent tasks run in parallel for faster completionCác task độc lập chạy song song để hoàn thành nhanh hơn
  • Use teams for large projects with multiple modulesDùng teams cho dự án lớn với nhiều module

Practice

Test your understanding of this chapter

Quiz

What is the main benefit of multi-agent orchestration?

Lợi ích chính của multi-agent orchestration là gì?

True or False

All tasks in a team must be executed sequentially.

Tất cả task trong team phải được thực hiện tuần tự.

Code Challenge

Complete the command to create a new team

Hoàn thành lệnh để tạo team mới

/team  "project-name"
Quiz

Which agent specializes in finding security vulnerabilities?

Agent nào chuyên tìm lỗ hổng bảo mật?

← → to navigate chapters
Built: 4/8/2026, 12:01:11 PM