What problem does it solve?
This Skill solves the problem of designing reliable, cross-runtime multi-agent orchestration plans that turn ambiguous “agent swarm / multi-agent collaboration” requests into an executable workflow with clear roles, skill discovery, and quality checks.
Core Features & Use Cases
- Agent Teams Orchestration (6-Stage Workflow): Plan first, discover skills, assemble a team, run parallel work, apply quality gates, and deliver an execution report.
- Cross-Runtime Compatibility & Degradation: Maps an abstract orchestration contract to Claude Code, Codex, OpenClaw, and Cursor, while explicitly degrading to “main-thread staged execution” when team/tool support is missing.
- Skill Fallback Chain for Reliability: Performs local skill scanning, then searches via find-skills when available, and finally falls back to general subagents or staged execution—without skipping required stages.
Quick Start
Ask the AI to run /agent-teams-playbook to coordinate an agent team that plans, discovers the best available skills, executes in parallel, and returns a quality-gated final deliverable.