From darkroom
Orchestrates parallel Claude Code agents for 3+ independent workstreams like frontend/backend/tests splits or large-scale refactors.
npx claudepluginhub darkroomengineering/cc-settingsThis skill uses the workspace's default tool permissions.
Before starting work, create a marker: `mkdir -p ~/.claude/tmp && echo "teams" > ~/.claude/tmp/heavy-skill-active && date -u +"%Y-%m-%dT%H:%M:%SZ" >> ~/.claude/tmp/heavy-skill-active`
Coordinates multiple Claude Code instances as agent teams with shared tasks, inter-agent messaging, and management for parallel research, feature development, debugging, and cross-layer coordination.
Orchestrates multi-agent teams to decompose complex tasks, select optimal sub-agents like Explore or Plan, and execute parallel work across domains such as frontend and backend.
Coordinates multiple Claude Code sessions as lead + teammates with shared tasks, messaging, file locks for parallel repo development. Includes sizing, decomposition patterns.
Share bugs, ideas, or general feedback.
Before starting work, create a marker: mkdir -p ~/.claude/tmp && echo "teams" > ~/.claude/tmp/heavy-skill-active && date -u +"%Y-%m-%dT%H:%M:%SZ" >> ~/.claude/tmp/heavy-skill-active
Delegates to the Maestro agent for coordinating multiple independent Claude Code instances.
| Scenario | Teams? | Why |
|---|---|---|
| 3+ independent file areas | Yes | Maximum parallelism |
| Frontend + Backend + Tests | Yes | No file conflicts |
| Sequential dependent work | No | Use subagents instead |
| Quick single investigation | No | Overhead not worth it |
| Large codebase analysis | Yes | Independent context per agent |
| Competing approaches | Yes | Explore in parallel |
CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS: "1" must be set (enabled by default in cc-settings)Report: team composition, task assignments, coordination strategy, and progress.