MCP HubMCP Hub
スキル一覧に戻る

Swarm Orchestration

proffesor-for-testing
更新日 Yesterday
99 閲覧
99
21
99
GitHubで表示
メタaiautomationdesign

について

Swarm Orchestrationは、開発者がメッシュや階層構造などの動的トポロジーを使用して、並列タスク実行のためのマルチエージェントAIシステムを調整できるようにします。これは、単一エージェントを超える複雑なワークフローのスケーリングに最適で、自動タスク分散、負荷分散、およびフォールトトレランスを備えています。専門化されたエージェント間でインテリジェントな調整を必要とする分散AIシステムを構築する際に、このスキルを使用してください。

クイックインストール

Claude Code

推奨
プラグインコマンド推奨
/plugin add https://github.com/proffesor-for-testing/agentic-qe
Git クローン代替
git clone https://github.com/proffesor-for-testing/agentic-qe.git ~/.claude/skills/Swarm Orchestration

このコマンドをClaude Codeにコピー&ペーストしてスキルをインストールします

ドキュメント

Swarm Orchestration

What This Skill Does

Orchestrates multi-agent swarms using agentic-flow's advanced coordination system. Supports mesh, hierarchical, and adaptive topologies with automatic task distribution, load balancing, and fault tolerance.

Prerequisites

  • agentic-flow v1.5.11+
  • Node.js 18+
  • Understanding of distributed systems (helpful)

Quick Start

# Initialize swarm
npx agentic-flow hooks swarm-init --topology mesh --max-agents 5

# Spawn agents
npx agentic-flow hooks agent-spawn --type coder
npx agentic-flow hooks agent-spawn --type tester
npx agentic-flow hooks agent-spawn --type reviewer

# Orchestrate task
npx agentic-flow hooks task-orchestrate \
  --task "Build REST API with tests" \
  --mode parallel

Topology Patterns

1. Mesh (Peer-to-Peer)

// Equal peers, distributed decision-making
await swarm.init({
  topology: 'mesh',
  agents: ['coder', 'tester', 'reviewer'],
  communication: 'broadcast'
});

2. Hierarchical (Queen-Worker)

// Centralized coordination, specialized workers
await swarm.init({
  topology: 'hierarchical',
  queen: 'architect',
  workers: ['backend-dev', 'frontend-dev', 'db-designer']
});

3. Adaptive (Dynamic)

// Automatically switches topology based on task
await swarm.init({
  topology: 'adaptive',
  optimization: 'task-complexity'
});

Task Orchestration

Parallel Execution

// Execute tasks concurrently
const results = await swarm.execute({
  tasks: [
    { agent: 'coder', task: 'Implement API endpoints' },
    { agent: 'frontend', task: 'Build UI components' },
    { agent: 'tester', task: 'Write test suite' }
  ],
  mode: 'parallel',
  timeout: 300000 // 5 minutes
});

Pipeline Execution

// Sequential pipeline with dependencies
await swarm.pipeline([
  { stage: 'design', agent: 'architect' },
  { stage: 'implement', agent: 'coder', after: 'design' },
  { stage: 'test', agent: 'tester', after: 'implement' },
  { stage: 'review', agent: 'reviewer', after: 'test' }
]);

Adaptive Execution

// Let swarm decide execution strategy
await swarm.autoOrchestrate({
  goal: 'Build production-ready API',
  constraints: {
    maxTime: 3600,
    maxAgents: 8,
    quality: 'high'
  }
});

Memory Coordination

// Share state across swarm
await swarm.memory.store('api-schema', {
  endpoints: [...],
  models: [...]
});

// Agents read shared memory
const schema = await swarm.memory.retrieve('api-schema');

Advanced Features

Load Balancing

// Automatic work distribution
await swarm.enableLoadBalancing({
  strategy: 'dynamic',
  metrics: ['cpu', 'memory', 'task-queue']
});

Fault Tolerance

// Handle agent failures
await swarm.setResiliency({
  retry: { maxAttempts: 3, backoff: 'exponential' },
  fallback: 'reassign-task'
});

Performance Monitoring

// Track swarm metrics
const metrics = await swarm.getMetrics();
// { throughput, latency, success_rate, agent_utilization }

Integration with Hooks

# Pre-task coordination
npx agentic-flow hooks pre-task --description "Build API"

# Post-task synchronization
npx agentic-flow hooks post-task --task-id "task-123"

# Session restore
npx agentic-flow hooks session-restore --session-id "swarm-001"

Best Practices

  1. Start small: Begin with 2-3 agents, scale up
  2. Use memory: Share context through swarm memory
  3. Monitor metrics: Track performance and bottlenecks
  4. Enable hooks: Automatic coordination and sync
  5. Set timeouts: Prevent hung tasks

Troubleshooting

Issue: Agents not coordinating

Solution: Verify memory access and enable hooks

Issue: Poor performance

Solution: Check topology (use adaptive) and enable load balancing

Learn More

  • Swarm Guide: docs/swarm/orchestration.md
  • Topology Patterns: docs/swarm/topologies.md
  • Hooks Integration: docs/hooks/coordination.md

GitHub リポジトリ

proffesor-for-testing/agentic-qe
パス: .claude/skills/swarm-orchestration
agenticqeagenticsfoundationagentsquality-engineering

関連スキル

content-collections

メタ

This skill provides a production-tested setup for Content Collections, a TypeScript-first tool that transforms Markdown/MDX files into type-safe data collections with Zod validation. Use it when building blogs, documentation sites, or content-heavy Vite + React applications to ensure type safety and automatic content validation. It covers everything from Vite plugin configuration and MDX compilation to deployment optimization and schema validation.

スキルを見る

creating-opencode-plugins

メタ

This skill provides the structure and API specifications for creating OpenCode plugins that hook into 25+ event types like commands, files, and LSP operations. It offers implementation patterns for JavaScript/TypeScript modules that intercept and extend the AI assistant's lifecycle. Use it when you need to build event-driven plugins for monitoring, custom handling, or extending OpenCode's capabilities.

スキルを見る

evaluating-llms-harness

テスト

This Claude Skill runs the lm-evaluation-harness to benchmark LLMs across 60+ standardized academic tasks like MMLU and GSM8K. It's designed for developers to compare model quality, track training progress, or report academic results. The tool supports various backends including HuggingFace and vLLM models.

スキルを見る

sglang

メタ

SGLang is a high-performance LLM serving framework that specializes in fast, structured generation for JSON, regex, and agentic workflows using its RadixAttention prefix caching. It delivers significantly faster inference, especially for tasks with repeated prefixes, making it ideal for complex, structured outputs and multi-turn conversations. Choose SGLang over alternatives like vLLM when you need constrained decoding or are building applications with extensive prefix sharing.

スキルを見る