Orchestrate teams of Claude Code sessions - coordinate multiple instances working together as a team with shared tasks and messaging
Coordinate multiple Claude Code instances working together as a team, with shared tasks, inter-agent messaging, and centralized management.
[!IMPORTANT] Agent teams are experimental and disabled by default. Enable them by adding
CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMSto yoursettings.jsonor environment. Agent teams have known limitations around session resumption, task coordination, and shutdown behavior.
Agent teams let you coordinate multiple Claude Code instances working together. One session acts as the team lead, coordinating work, assigning tasks, and synthesizing results. Teammates work independently, each in its own context window, and communicate directly with each other.
Unlike subagents, which run within a single session and can only report back to the main agent, you can also interact with individual teammates directly without going through the lead.
[!TIP]
Agent teams require Claude Code v2.1.32 or later. Check your version with claude --version.
Agent teams are most effective for tasks where parallel exploration adds real value:
| Feature | Subagents | Agent teams |
|---|---|---|
| Context | Results return to caller | Fully independent |
| Communication | Report back to lead only | Teammates message each other |
| Coordination | Lead manages all work | Shared task list |
Add the environment variable CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS to your shell or ~/.claude.json:
{
"env": {
"CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS": "1"
}
}
Ask Claude to create a team describing the task and structure: "I'm designing a CLI tool... Create an agent team to explore this from different angles: one on UX, one on architecture, one as devil's advocate."
The complete documentation index is available at: llms.txt Use this index to discover other available pages.