swarm-orchestration

star 0

Multi-agent swarm orchestration patterns from kyegomez/swarms. Sequential, Concurrent, MoA, GroupChat, Hierarchical, and SwarmRouter architectures for coordinating multiple agents.

Geargrindadmin By Geargrindadmin schedule Updated 3/7/2026

name: swarm-orchestration description: Multi-agent swarm orchestration patterns from kyegomez/swarms. Sequential, Concurrent, MoA, GroupChat, Hierarchical, and SwarmRouter architectures for coordinating multiple agents. trigger: multi-agent orchestration, swarm coordination, agent workflow, parallel agents, sequential workflow, agent communication

Swarm Orchestration Patterns

Reference patterns extracted from kyegomez/swarms — The Enterprise-Grade Multi-Agent Orchestration Framework.

Architecture Patterns

SequentialWorkflow

Agents execute tasks in linear order, each building on the previous output.

  • Use when: Tasks have strict dependencies
  • Pattern: Agent A → Agent B → Agent C

ConcurrentWorkflow

Agents execute tasks in parallel, results aggregated.

  • Use when: Independent tasks can run simultaneously
  • Pattern: Fan-out → parallel execution → fan-in

AgentRearrange

Dynamic agent routing based on a topology string (e.g., "A → B, A → C, B → D, C → D").

  • Use when: Complex DAG workflows with conditional branching

SwarmRouter (Universal Orchestrator)

Routes tasks to the optimal swarm type automatically.

  • Use when: You want automatic orchestration selection based on task characteristics

MixtureOfAgents (MoA)

Multiple agents provide independent assessments, synthesized by a coordinator.

  • Use when: Decisions require diverse perspectives (board meetings, peer review)

GroupChat

Agents converse in a shared context, building on each other's contributions.

  • Use when: Collaborative problem-solving, brainstorming

HierarchicalSwarm

Tree-structured agent hierarchy. Manager agents delegate to specialist sub-agents.

  • Use when: Large-scale tasks with clear domain separation

HeavySwarm

All agents process full context and vote on outputs.

  • Use when: Critical decisions requiring consensus

Communication Protocols

  • Agent-to-Agent: Direct message passing with typed interfaces
  • Broadcast: One-to-many task distribution
  • Coordinator Pattern: Central coordinator manages agent lifecycle
  • AOP (Agent Orchestration Protocol): Standardized agent discovery and communication

GGV3 Application

For GearGrind, use these patterns in the existing Agent Orchestration system:

  • Board meetings → MixtureOfAgents (CA, CPO, CSO, COO, CXO)
  • Implementation → SequentialWorkflow (Research → Plan → Implement → Test → Ship)
  • Multi-domain tasks → HierarchicalSwarm (Coordinator → specialist agents)
  • Code review → ConcurrentWorkflow (Security + Quality + Design agents in parallel)
Install via CLI
npx skills add https://github.com/Geargrindadmin/gg-agentic-harness --skill swarm-orchestration
Repository Details
star Stars 0
call_split Forks 0
navigation Branch main
article Path SKILL.md
More from Creator
Geargrindadmin
Geargrindadmin Explore all skills →