swarm-orchestration
Scannednpx machina-cli add skill a5c-ai/babysitter/swarm-orchestration --openclawSwarm Orchestration
Overview
Form and coordinate multi-agent swarms with topology-aware deployment. Supports Mesh, Hierarchical, Ring, and Star topologies with automatic selection based on task complexity and agent count.
When to Use
- Complex tasks requiring multiple specialized agents
- Tasks needing coordinated parallel execution
- When consensus among agents is required for quality
- Projects requiring anti-drift enforcement during execution
Process
- Topology Selection - Analyze task and agent pool to select optimal topology
- Agent Assignment - Assign Queen (Strategic/Tactical/Adaptive) and Worker roles
- Consensus Init - Initialize Raft/Byzantine/Gossip/CRDT protocol
- Parallel Execution - Distribute subtasks with shared memory
- Anti-Drift Checkpoints - Validate alignment every N subtasks
- Consensus Voting - Weighted voting (Queen=3x) for final decision
Topologies
- Mesh: All-to-all communication, best for small swarms (<8 agents)
- Hierarchical: Queen coordinates workers, best for large/structured tasks
- Ring: Sequential handoff, best for pipeline/transformation tasks
- Star: Central coordinator fan-out, best for independent subtasks
Agents Used
agents/strategic-queen/- Long-term planning swarmsagents/tactical-queen/- Execution coordination swarmsagents/adaptive-queen/- Real-time optimization swarmsagents/swarm-coordinator/- Topology management
Tool Use
Invoke via babysitter process: methodologies/ruflo/ruflo-swarm-coordination
Source
git clone https://github.com/a5c-ai/babysitter/blob/main/plugins/babysitter/skills/babysit/process/methodologies/ruflo/skills/swarm-orchestration/SKILL.mdView on GitHub Overview
Swarm orchestration forms and coordinates multi-agent swarms with topology-aware deployment and automatic topology selection. It supports Mesh, Hierarchical, Ring, and Star layouts and assigns Queen and Worker roles to optimize task execution while enforcing anti-drift and coordinated consensus.
How This Skill Works
The system analyzes the task and agent pool to select an optimal topology, then assigns Queen and Worker roles. It initializes a consensus protocol (Raft/Byzantine/Gossip/CRDT), distributes subtasks for parallel execution, and runs anti-drift checkpoints with weighted voting for the final decision.
When to Use It
- When a task requires multiple specialized agents to operate in parallel.
- When consensus among agents is critical for output quality.
- When anti-drift enforcement is needed during long-running tasks.
- When task complexity benefits from topology-aware deployment.
- When coordinating execution across diverse agent roles in large pools.
Quick Start
- Step 1: Invoke the swarm coordination via babysitter: methodologies/ruflo/ruflo-swarm-coordination
- Step 2: Allow the system to select topology and assign Queen/Worker roles
- Step 3: Run the parallel tasks with anti-drift checkpoints and consensus voting
Best Practices
- Define clear Queen roles (strategic, tactical, adaptive) upfront.
- Choose topology based on task structure and swarm size; mix topologies when appropriate.
- Initialize and test the chosen consensus protocol before execution.
- Implement anti-drift checkpoints at regular subtasks.
- Utilize weighted voting (Queen = 3x) to reach robust final decisions.
Example Use Cases
- A 6-agent swarm processes real-time video frames using Ring topology for pipeline-like steps.
- Hierarchical deployment coordinates planning and execution for a large data-analysis task.
- Mesh topology aggregates data from several small swarms for quick, parallel data collection.
- Star topology distributes independent subtasks under a central coordinator for efficiency.
- Consensus-driven reallocation of tasks when a worker fails to maintain progress.
