From enforce-teams
Configure Enforce Teams delegation level — controls when the team leader automatically creates Agent Teams
How this command is triggered — by the user, by Claude, or both
Slash command
/enforce-teams:enforce-teams Optional level (full, smart, off)The summary Claude sees in its command listing — used to decide when to auto-load this command
# Enforce Teams Configuration You are helping the user configure the Enforce Teams plugin, which controls when the team leader automatically creates Agent Teams (orchestrate teams of Claude Code sessions). ## Step 0: Prerequisites Check Check if the environment variable `CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS` is set to `1` or `true`. Run: If it is NOT set (empty or missing), warn the user: > Agent Teams requires the experimental flag. Add this to your shell profile: > > Then restart your terminal and try again. Stop here if the flag is not set. ## Step 1: Read Current Level Read t...
You are helping the user configure the Enforce Teams plugin, which controls when the team leader automatically creates Agent Teams (orchestrate teams of Claude Code sessions).
Check if the environment variable CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS is set to 1 or true. Run:
echo $CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS
If it is NOT set (empty or missing), warn the user:
Agent Teams requires the experimental flag. Add this to your shell profile:
export CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1Then restart your terminal and try again.
Stop here if the flag is not set.
Read the file ~/.claude/enforce-teams-level. It contains a single word: full, smart, or off.
smart.If $ARGUMENTS is provided and matches a valid level name (full, smart, off), skip asking and use that level directly. Go to Step 3.
Otherwise, display the following level table to the user. Mark the current level with "← Current". Then ask which level they want.
┌───┬──────────────────┬──────────────────────────────────────────────────────────┐
│ # │ Level │ Behavior │
├───┼──────────────────┼──────────────────────────────────────────────────────────┤
│ 3 │ Full Delegate │ Pure coordinator. Delegates ALL work to teammates. │
│ 2 │ Smart Delegate ★ │ Creates teams for complex tasks. Handles simple directly.│
│ 1 │ Off │ Default Claude Code behavior. Teams require your approval│
└───┴──────────────────┴──────────────────────────────────────────────────────────┘
★ = Recommended
Use AskUserQuestion to let the user pick:
{
"questions": [
{
"question": "Select delegation level:",
"header": "Enforce Teams Level",
"multiSelect": false,
"options": [
{
"label": "3 - Full Delegate",
"description": "Pure coordinator. Delegates ALL work to teammates — never reads, writes, or searches files directly."
},
{
"label": "2 - Smart Delegate (Recommended)",
"description": "Creates teams for complex tasks (3+ files, research, debug, review). Handles simple stuff directly."
},
{
"label": "1 - Off",
"description": "Restores default Claude Code behavior. Claude may suggest teams but always asks for your approval first."
}
]
}
]
}
Write the chosen level keyword to ~/.claude/enforce-teams-level:
fullsmartoffecho "full" > ~/.claude/enforce-teams-level
(Replace full with the appropriate level keyword.)
After writing, tell the user:
/enforce-teams or /enforce-teams <level>Examples per level:
npx claudepluginhub wan-zl/claude-enforce-teams --plugin enforce-teams/team-spawnSpawns a multi-agent team from presets (review, debug, feature, fullstack, research, security, migration) or custom composition with optional name, member count, and delegation mode.
/teamManages Claude Code Agent Teams for multi-agent coordination via shared tasks, messaging, and parallel execution. Supports create, status, wave, review, research, build, clean subcommands.
/codeLaunches a code-mode agent team with configurable tier and research settings. Guides through outcomes, team setup, confirmation, and launch.
/team-claudeCreates a Claude agent team with Opus orchestrating Sonnet/Haiku workers — spawns teammates, assigns tasks, coordinates execution, and shuts down on completion.
/criar-teamGuides creation of an Agent Team using two architectures: v2.8 parallel subagents (file-based consolidation) or v2.10 real-time debate (experimental). Asks discovery questions, reads specs, and generates code.
/teamAssembles, configures, and spawns native Claude Code agent teams for parallel tasks, code reviews, full-stack development, and security audits. Supports create, context, list subcommands.