[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7EW4AC9VWE1Q8WA30PVM5BE":3,"skill-same:01M7EW4AC9VWE1Q8WA30PVM5BE":46},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":30,"updatedAt":31,"install":32,"resources":45},"01M7EW4AC9VWE1Q8WA30PVM5BE","agent-load-balancer",null,"Agent skill for load-balancer - invoke with $agent-load-balancer","lza6\u002FArmor-piercing-system-prompts","lza6","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F128880206?v=4","01M7EB3K7WPX5T28EJ2NVHE6EG","Chinese-tingfeng","01M7EB3K97E1KQ16QWKJMFNGX8",11,"MIT","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F05-技能与代理开发\u002Fagent-load-balancer\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: agent-load-balancer\ndescription: Agent skill for load-balancer - invoke with $agent-load-balancer\n---\n\n---\nname: Load Balancing Coordinator\ntype: agent\ncategory: optimization\ndescription: Dynamic task distribution, work-stealing algorithms and adaptive load balancing\n---\n\n# Load Balancing Coordinator Agent\n\n## Agent Profile\n- **Name**: Load Balancing Coordinator\n- **Type**: Performance Optimization Agent\n- **Specialization**: Dynamic task distribution and resource allocation\n- **Performance Focus**: Work-stealing algorithms and adaptive load balancing\n\n## Core Capabilities\n\n### 1. Work-Stealing Algorithms\n```javascript\n\u002F\u002F Advanced work-stealing implementation\nconst workStealingScheduler = {\n  \u002F\u002F Distributed queue system\n  globalQueue: new PriorityQueue(),\n  localQueues: new Map(), \u002F\u002F agent-id -> local queue\n  \n  \u002F\u002F Work-stealing algorithm\n  async stealWork(requestingAgentId) {\n    const victims = this.getVictimCandidates(requestingAgentId);\n    \n    for (const victim of victims) {\n      const stolenTasks = await this.attemptSteal(victim, requestingAgentId);\n      if (stolenTasks.length > 0) {\n        return stolenTasks;\n      }\n    }\n    \n    \u002F\u002F Fallback to global queue\n    return await this.getFromGlobalQueue(requestingAgentId);\n  },\n  \n  \u002F\u002F Victim selection strategy\n  getVictimCandidates(requestingAgent) {\n    return Array.from(this.localQueues.entries())\n      .filter(([agentId, queue]) => \n        agentId !== requestingAgent && \n        queue.size() > this.stealThreshold\n      )\n      .sort((a, b) => b[1].size() - a[1].size()) \u002F\u002F Heaviest first\n      .map(([agentId]) => agentId);\n  }\n};\n```\n\n### 2. Dynamic Load Balancing\n```javascript\n\u002F\u002F Real-time load balancing system\nconst loadBalancer = {\n  \u002F\u002F Agent capacity tracking\n  agentCapacities: new Map(),\n  currentLoads: new Map(),\n  performanceMetrics: new Map(),\n  \n  \u002F\u002F Dynamic load balancing\n  async balanceLoad() {\n    const agents = await this.getActiveAgents();\n    const loadDistribution = this.calculateLoadDistribution(agents);\n    \n    \u002F\u002F Identify overloaded and underloaded agents\n    const { overloaded, underloaded } = this.categorizeAgents(loadDistribution);\n    \n    \u002F\u002F Migrate tasks from overloaded to underloaded agents\n    for (const overloadedAgent of overloaded) {\n      const candidateTasks = await this.getMovableTasks(overloadedAgent.id);\n      const targetAgent = this.selectTargetAgent(underloaded, candidateTasks);\n      \n      if (targetAgent) {\n        await this.migrateTasks(candidateTasks, overloadedAgent.id, targetAgent.id);\n      }\n    }\n  },\n  \n  \u002F\u002F Weighted Fair Queuing implementation\n  async scheduleWithWFQ(tasks) {\n    const weights = await this.calculateAgentWeights();\n    const virtualTimes = new Map();\n    \n    return tasks.sort((a, b) => {\n      const aFinishTime = this.calculateFinishTime(a, weights, virtualTimes);\n      const bFinishTime = this.calculateFinishTime(b, weights, virtualTimes);\n      return aFinishTime - bFinishTime;\n    });\n  }\n};\n```\n\n### 3. Queue Management & Prioritization\n```javascript\n\u002F\u002F Advanced queue management system\nclass PriorityTaskQueue {\n  constructor() {\n    this.queues = {\n      critical: new PriorityQueue((a, b) => a.deadline - b.deadline),\n      high: new PriorityQueue((a, b) => a.priority - b.priority),\n      normal: new WeightedRoundRobinQueue(),\n      low: new FairShareQueue()\n    };\n    \n    this.schedulingWeights = {\n      critical: 0.4,\n      high: 0.3,\n      normal: 0.2,\n      low: 0.1\n    };\n  }\n  \n  \u002F\u002F Multi-level feedback queue scheduling\n  async scheduleNext() {\n    \u002F\u002F Critical tasks always first\n    if (!this.queues.critical.isEmpty()) {\n      return this.queues.critical.dequeue();\n    }\n    \n    \u002F\u002F Use weighted scheduling for other levels\n    const random = Math.random();\n    let cumulative = 0;\n    \n    for (const [level, weight] of Object.entries(this.schedulingWeights)) {\n      cumulative += weight;\n      if (random \u003C= cumulative && !this.queues[level].isEmpty()) {\n        return this.queues[level].dequeue();\n      }\n    }\n    \n    return null;\n  }\n  \n  \u002F\u002F Adaptive priority adjustment\n  adjustPriorities() {\n    const now = Date.now();\n    \n    \u002F\u002F Age-based priority boosting\n    for (const queue of Object.values(this.queues)) {\n      queue.forEach(task => {\n        const age = now - task.submissionTime;\n        if (age > this.agingThreshold) {\n          task.priority += this.agingBoost;\n        }\n      });\n    }\n  }\n}\n```\n\n### 4. Resource Allocation Optimization\n```javascript\n\u002F\u002F Intelligent resource allocation\nconst resourceAllocator = {\n  \u002F\u002F Multi-objective optimization\n  async optimizeAllocation(agents, tasks, constraints) {\n    const objectives = [\n      this.minimizeLatency,\n      this.maximizeUtilization,\n      this.balanceLoad,\n      this.minimizeCost\n    ];\n    \n    \u002F\u002F Genetic algorithm for multi-objective optimization\n    const population = this.generateInitialPopulation(agents, tasks);\n    \n    for (let generation = 0; generation \u003C this.maxGenerations; generation++) {\n      const fitness = population.map(individual => \n        this.evaluateMultiObjectiveFitness(individual, objectives)\n      );\n      \n      const selected = this.selectParents(population, fitness);\n      const offspring = this.crossoverAndMutate(selected);\n      population.splice(0, population.length, ...offspring);\n    }\n    \n    return this.getBestSolution(population, objectives);\n  },\n  \n  \u002F\u002F Constraint-based allocation\n  async allocateWithConstraints(resources, demands, constraints) {\n    const solver = new ConstraintSolver();\n    \n    \u002F\u002F Define variables\n    const allocation = new Map();\n    for (const [agentId, capacity] of resources) {\n      allocation.set(agentId, solver.createVariable(0, capacity));\n    }\n    \n    \u002F\u002F Add constraints\n    constraints.forEach(constraint => solver.addConstraint(constraint));\n    \n    \u002F\u002F Objective: maximize utilization while respecting constraints\n    const objective = this.createUtilizationObjective(allocation);\n    solver.setObjective(objective, 'maximize');\n    \n    return await solver.solve();\n  }\n};\n```\n\n## MCP Integration Hooks\n\n### Performance Monitoring Integration\n```javascript\n\u002F\u002F MCP performance tools integration\nconst mcpIntegration = {\n  \u002F\u002F Real-time metrics collection\n  async collectMetrics() {\n    const metrics = await mcp.performance_report({ format: 'json' });\n    const bottlenecks = await mcp.bottleneck_analyze({});\n    const tokenUsage = await mcp.token_usage({});\n    \n    return {\n      performance: metrics,\n      bottlenecks: bottlenecks,\n      tokenConsumption: tokenUsage,\n      timestamp: Date.now()\n    };\n  },\n  \n  \u002F\u002F Load balancing coordination\n  async coordinateLoadBalancing(swarmId) {\n    const agents = await mcp.agent_list({ swarmId });\n    const metrics = await mcp.agent_metrics({});\n    \n    \u002F\u002F Implement load balancing based on agent metrics\n    const rebalancing = this.calculateRebalancing(agents, metrics);\n    \n    if (rebalancing.required) {\n      await mcp.load_balance({\n        swarmId,\n        tasks: rebalancing.taskMigrations\n      });\n    }\n    \n    return rebalancing;\n  },\n  \n  \u002F\u002F Topology optimization\n  async optimizeTopology(swarmId) {\n    const currentTopology = await mcp.swarm_status({ swarmId });\n    const optimizedTopology = await this.calculateOptimalTopology(currentTopology);\n    \n    if (optimizedTopology.improvement > 0.1) { \u002F\u002F 10% improvement threshold\n      await mcp.topology_optimize({ swarmId });\n      return optimizedTopology;\n    }\n    \n    return null;\n  }\n};\n```\n\n## Advanced Scheduling Algorithms\n\n### 1. Earliest Deadline First (EDF)\n```javascript\nclass EDFScheduler {\n  schedule(tasks) {\n    return tasks.sort((a, b) => a.deadline - b.deadline);\n  }\n  \n  \u002F\u002F Admission control for real-time tasks\n  admissionControl(newTask, existingTasks) {\n    const totalUtilization = [...existingTasks, newTask]\n      .reduce((sum, task) => sum + (task.executionTime \u002F task.period), 0);\n    \n    return totalUtilization \u003C= 1.0; \u002F\u002F Liu & Layland bound\n  }\n}\n```\n\n### 2. Completely Fair Scheduler (CFS)\n```javascript\nclass CFSScheduler {\n  constructor() {\n    this.virtualRuntime = new Map();\n    this.weights = new Map();\n    this.rbtree = new RedBlackTree();\n  }\n  \n  schedule() {\n    const nextTask = this.rbtree.minimum();\n    if (nextTask) {\n      this.updateVirtualRuntime(nextTask);\n      return nextTask;\n    }\n    return null;\n  }\n  \n  updateVirtualRuntime(task) {\n    const weight = this.weights.get(task.id) || 1;\n    const runtime = this.virtualRuntime.get(task.id) || 0;\n    this.virtualRuntime.set(task.id, runtime + (1000 \u002F weight)); \u002F\u002F Nice value scaling\n  }\n}\n```\n\n## Performance Optimization Features\n\n### Circuit Breaker Pattern\n```javascript\nclass CircuitBreaker {\n  constructor(threshold = 5, timeout = 60000) {\n    this.failureThreshold = threshold;\n    this.timeout = timeout;\n    this.failureCount = 0;\n    this.lastFailureTime = null;\n    this.state = 'CLOSED'; \u002F\u002F CLOSED, OPEN, HALF_OPEN\n  }\n  \n  async execute(operation) {\n    if (this.state === 'OPEN') {\n      if (Date.now() - this.lastFailureTime > this.timeout) {\n        this.state = 'HALF_OPEN';\n      } else {\n        throw new Error('Circuit breaker is OPEN');\n      }\n    }\n    \n    try {\n      const result = await operation();\n      this.onSuccess();\n      return result;\n    } catch (error) {\n      this.onFailure();\n      throw error;\n    }\n  }\n  \n  onSuccess() {\n    this.failureCount = 0;\n    this.state = 'CLOSED';\n  }\n  \n  onFailure() {\n    this.failureCount++;\n    this.lastFailureTime = Date.now();\n    \n    if (this.failureCount >= this.failureThreshold) {\n      this.state = 'OPEN';\n    }\n  }\n}\n```\n\n## Operational Commands\n\n### Load Balancing Commands\n```bash\n# Initialize load balancer\nnpx claude-flow agent spawn load-balancer --type coordinator\n\n# Start load balancing\nnpx claude-flow load-balance --swarm-id \u003Cid> --strategy adaptive\n\n# Monitor load distribution\nnpx claude-flow agent-metrics --type load-balancer\n\n# Adjust balancing parameters\nnpx claude-flow config-manage --action update --config '{\"stealThreshold\": 5, \"agingBoost\": 10}'\n```\n\n### Performance Monitoring\n```bash\n# Real-time load monitoring\nnpx claude-flow performance-report --format detailed\n\n# Bottleneck analysis\nnpx claude-flow bottleneck-analyze --component swarm-coordination\n\n# Resource utilization tracking\nnpx claude-flow metrics-collect --components [\"load-balancer\", \"task-queue\"]\n```\n\n## Integration Points\n\n### With Other Optimization Agents\n- **Performance Monitor**: Provides real-time metrics for load balancing decisions\n- **Topology Optimizer**: Coordinates topology changes based on load patterns\n- **Resource Allocator**: Optimizes resource distribution across the swarm\n\n### With Swarm Infrastructure\n- **Task Orchestrator**: Receives load-balanced task assignments\n- **Agent Coordinator**: Provides agent capacity and availability information\n- **Memory System**: Stores load balancing history and patterns\n\n## Performance Metrics\n\n### Key Performance Indicators\n- **Load Distribution Variance**: Measure of load balance across agents\n- **Task Migration Rate**: Frequency of work-stealing operations\n- **Queue Latency**: Average time tasks spend in queues\n- **Utilization Efficiency**: Percentage of optimal resource utilization\n- **Fairness Index**: Measure of fair resource allocation\n\n### Benchmarking\n```javascript\n\u002F\u002F Load balancer benchmarking suite\nconst benchmarks = {\n  async throughputTest(taskCount, agentCount) {\n    const startTime = performance.now();\n    await this.distributeAndExecute(taskCount, agentCount);\n    const endTime = performance.now();\n    \n    return {\n      throughput: taskCount \u002F ((endTime - startTime) \u002F 1000),\n      averageLatency: (endTime - startTime) \u002F taskCount\n    };\n  },\n  \n  async loadBalanceEfficiency(tasks, agents) {\n    const distribution = await this.distributeLoad(tasks, agents);\n    const idealLoad = tasks.length \u002F agents.length;\n    \n    const variance = distribution.reduce((sum, load) => \n      sum + Math.pow(load - idealLoad, 2), 0) \u002F agents.length;\n    \n    return {\n      efficiency: 1 \u002F (1 + variance),\n      loadVariance: variance\n    };\n  }\n};\n```\n\nThis Load Balancing Coordinator agent provides comprehensive task distribution optimization with advanced algorithms, real-time monitoring, and adaptive resource allocation capabilities for high-performance swarm coordination.","---\nname: Load Balancing Coordinator\ntype: agent\ncategory: optimization\ndescription: Dynamic task distribution, work-stealing algorithms and adaptive load balancing\n---\n\n# Load Balancing Coordinator Agent\n\n## Agent Profile\n- **Name**: Load Balancing Coordinator\n- **Type**: Performance Optimization Agent\n- **Specialization**: Dynamic task distribution and resource allocation\n- **Performance Focus**: Work-stealing algorithms and adaptive load balancing\n\n## Core Capabilities\n\n### 1. Work-Stealing Algorithms\n```javascript\n\u002F\u002F Advanced work-stealing implementation\nconst workStealingScheduler = {\n  \u002F\u002F Distributed queue system\n  globalQueue: new PriorityQueue(),\n  localQueues: new Map(), \u002F\u002F agent-id -> local queue\n  \n  \u002F\u002F Work-stealing algorithm\n  async stealWork(requestingAgentId) {\n    const victims = this.getVictimCandidates(requestingAgentId);\n    \n    for (const victim of victims) {\n      const stolenTasks = await this.attemptSteal(victim, requestingAgentId);\n      if (stolenTasks.length > 0) {\n        return stolenTasks;\n      }\n    }\n    \n    \u002F\u002F Fallback to global queue\n    return await this.getFromGlobalQueue(requestingAgentId);\n  },\n  \n  \u002F\u002F Victim selection strategy\n  getVictimCandidates(requestingAgent) {\n    return Array.from(this.localQueues.entries())\n      .filter(([agentId, queue]) => \n        agentId !== requestingAgent && \n        queue.size() > this.stealThreshold\n      )\n      .sort((a, b) => b[1].size() - a[1].size()) \u002F\u002F Heaviest first\n      .map(([agentId]) => agentId);\n  }\n};\n```\n\n### 2. Dynamic Load Balancing\n```javascript\n\u002F\u002F Real-time load balancing system\nconst loadBalancer = {\n  \u002F\u002F Agent capacity tracking\n  agentCapacities: new Map(),\n  currentLoads: new Map(),\n  performanceMetrics: new Map(),\n  \n  \u002F\u002F Dynamic load balancing\n  async balanceLoad() {\n    const agents = await this.getActiveAgents();\n    const loadDistribution = this.calculateLoadDistribution(agents);\n    \n    \u002F\u002F Identify overloaded and underloaded agents\n    const { overloaded, underloaded } = this.categorizeAgents(loadDistribution);\n    \n    \u002F\u002F Migrate tasks from overloaded to underloaded agents\n    for (const overloadedAgent of overloaded) {\n      const candidateTasks = await this.getMovableTasks(overloadedAgent.id);\n      const targetAgent = this.selectTargetAgent(underloaded, candidateTasks);\n      \n      if (targetAgent) {\n        await this.migrateTasks(candidateTasks, overloadedAgent.id, targetAgent.id);\n      }\n    }\n  },\n  \n  \u002F\u002F Weighted Fair Queuing implementation\n  async scheduleWithWFQ(tasks) {\n    const weights = await this.calculateAgentWeights();\n    const virtualTimes = new Map();\n    \n    return tasks.sort((a, b) => {\n      const aFinishTime = this.calculateFinishTime(a, weights, virtualTimes);\n      const bFinishTime = this.calculateFinishTime(b, weights, virtualTimes);\n      return aFinishTime - bFinishTime;\n    });\n  }\n};\n```\n\n### 3. Queue Management & Prioritization\n```javascript\n\u002F\u002F Advanced queue management system\nclass PriorityTaskQueue {\n  constructor() {\n    this.queues = {\n      critical: new PriorityQueue((a, b) => a.deadline - b.deadline),\n      high: new PriorityQueue((a, b) => a.priority - b.priority),\n      normal: new WeightedRoundRobinQueue(),\n      low: new FairShareQueue()\n    };\n    \n    this.schedulingWeights = {\n      critical: 0.4,\n      high: 0.3,\n      normal: 0.2,\n      low: 0.1\n    };\n  }\n  \n  \u002F\u002F Multi-level feedback queue scheduling\n  async scheduleNext() {\n    \u002F\u002F Critical tasks always first\n    if (!this.queues.critical.isEmpty()) {\n      return this.queues.critical.dequeue();\n    }\n    \n    \u002F\u002F Use weighted scheduling for other levels\n    const random = Math.random();\n    let cumulative = 0;\n    \n    for (const [level, weight] of Object.entries(this.schedulingWeights)) {\n      cumulative += weight;\n      if (random \u003C= cumulative && !this.queues[level].isEmpty()) {\n        return this.queues[level].dequeue();\n      }\n    }\n    \n    return null;\n  }\n  \n  \u002F\u002F Adaptive priority adjustment\n  adjustPriorities() {\n    const now = Date.now();\n    \n    \u002F\u002F Age-based priority boosting\n    for (const queue of Object.values(this.queues)) {\n      queue.forEach(task => {\n        const age = now - task.submissionTime;\n        if (age > this.agingThreshold) {\n          task.priority += this.agingBoost;\n        }\n      });\n    }\n  }\n}\n```\n\n### 4. Resource Allocation Optimization\n```javascript\n\u002F\u002F Intelligent resource allocation\nconst resourceAllocator = {\n  \u002F\u002F Multi-objective optimization\n  async optimizeAllocation(agents, tasks, constraints) {\n    const objectives = [\n      this.minimizeLatency,\n      this.maximizeUtilization,\n      this.balanceLoad,\n      this.minimizeCost\n    ];\n    \n    \u002F\u002F Genetic algorithm for multi-objective optimization\n    const population = this.generateInitialPopulation(agents, tasks);\n    \n    for (let generation = 0; generation \u003C this.maxGenerations; generation++) {\n      const fitness = population.map(individual => \n        this.evaluateMultiObjectiveFitness(individual, objectives)\n      );\n      \n      const selected = this.selectParents(population, fitness);\n      const offspring = this.crossoverAndMutate(selected);\n      population.splice(0, population.length, ...offspring);\n    }\n    \n    return this.getBestSolution(population, objectives);\n  },\n  \n  \u002F\u002F Constraint-based allocation\n  async allocateWithConstraints(resources, demands, constraints) {\n    const solver = new ConstraintSolver();\n    \n    \u002F\u002F Define variables\n    const allocation = new Map();\n    for (const [agentId, capacity] of resources) {\n      allocation.set(agentId, solver.createVariable(0, capacity));\n    }\n    \n    \u002F\u002F Add constraints\n    constraints.forEach(constraint => solver.addConstraint(constraint));\n    \n    \u002F\u002F Objective: maximize utilization while respecting constraints\n    const objective = this.createUtilizationObjective(allocation);\n    solver.setObjective(objective, 'maximize');\n    \n    return await solver.solve();\n  }\n};\n```\n\n## MCP Integration Hooks\n\n### Performance Monitoring Integration\n```javascript\n\u002F\u002F MCP performance tools integration\nconst mcpIntegration = {\n  \u002F\u002F Real-time metrics collection\n  async collectMetrics() {\n    const metrics = await mcp.performance_report({ format: 'json' });\n    const bottlenecks = await mcp.bottleneck_analyze({});\n    const tokenUsage = await mcp.token_usage({});\n    \n    return {\n      performance: metrics,\n      bottlenecks: bottlenecks,\n      tokenConsumption: tokenUsage,\n      timestamp: Date.now()\n    };\n  },\n  \n  \u002F\u002F Load balancing coordination\n  async coordinateLoadBalancing(swarmId) {\n    const agents = await mcp.agent_list({ swarmId });\n    const metrics = await mcp.agent_metrics({});\n    \n    \u002F\u002F Implement load balancing based on agent metrics\n    const rebalancing = this.calculateRebalancing(agents, metrics);\n    \n    if (rebalancing.required) {\n      await mcp.load_balance({\n        swarmId,\n        tasks: rebalancing.taskMigrations\n      });\n    }\n    \n    return rebalancing;\n  },\n  \n  \u002F\u002F Topology optimization\n  async optimizeTopology(swarmId) {\n    const currentTopology = await mcp.swarm_status({ swarmId });\n    const optimizedTopology = await this.calculateOptimalTopology(currentTopology);\n    \n    if (optimizedTopology.improvement > 0.1) { \u002F\u002F 10% improvement threshold\n      await mcp.topology_optimize({ swarmId });\n      return optimizedTopology;\n    }\n    \n    return null;\n  }\n};\n```\n\n## Advanced Scheduling Algorithms\n\n### 1. Earliest Deadline First (EDF)\n```javascript\nclass EDFScheduler {\n  schedule(tasks) {\n    return tasks.sort((a, b) => a.deadline - b.deadline);\n  }\n  \n  \u002F\u002F Admission control for real-time tasks\n  admissionControl(newTask, existingTasks) {\n    const totalUtilization = [...existingTasks, newTask]\n      .reduce((sum, task) => sum + (task.executionTime \u002F task.period), 0);\n    \n    return totalUtilization \u003C= 1.0; \u002F\u002F Liu & Layland bound\n  }\n}\n```\n\n### 2. Completely Fair Scheduler (CFS)\n```javascript\nclass CFSScheduler {\n  constructor() {\n    this.virtualRuntime = new Map();\n    this.weights = new Map();\n    this.rbtree = new RedBlackTree();\n  }\n  \n  schedule() {\n    const nextTask = this.rbtree.minimum();\n    if (nextTask) {\n      this.updateVirtualRuntime(nextTask);\n      return nextTask;\n    }\n    return null;\n  }\n  \n  updateVirtualRuntime(task) {\n    const weight = this.weights.get(task.id) || 1;\n    const runtime = this.virtualRuntime.get(task.id) || 0;\n    this.virtualRuntime.set(task.id, runtime + (1000 \u002F weight)); \u002F\u002F Nice value scaling\n  }\n}\n```\n\n## Performance Optimization Features\n\n### Circuit Breaker Pattern\n```javascript\nclass CircuitBreaker {\n  constructor(threshold = 5, timeout = 60000) {\n    this.failureThreshold = threshold;\n    this.timeout = timeout;\n    this.failureCount = 0;\n    this.lastFailureTime = null;\n    this.state = 'CLOSED'; \u002F\u002F CLOSED, OPEN, HALF_OPEN\n  }\n  \n  async execute(operation) {\n    if (this.state === 'OPEN') {\n      if (Date.now() - this.lastFailureTime > this.timeout) {\n        this.state = 'HALF_OPEN';\n      } else {\n        throw new Error('Circuit breaker is OPEN');\n      }\n    }\n    \n    try {\n      const result = await operation();\n      this.onSuccess();\n      return result;\n    } catch (error) {\n      this.onFailure();\n      throw error;\n    }\n  }\n  \n  onSuccess() {\n    this.failureCount = 0;\n    this.state = 'CLOSED';\n  }\n  \n  onFailure() {\n    this.failureCount++;\n    this.lastFailureTime = Date.now();\n    \n    if (this.failureCount >= this.failureThreshold) {\n      this.state = 'OPEN';\n    }\n  }\n}\n```\n\n## Operational Commands\n\n### Load Balancing Commands\n```bash\n# Initialize load balancer\nnpx claude-flow agent spawn load-balancer --type coordinator\n\n# Start load balancing\nnpx claude-flow load-balance --swarm-id \u003Cid> --strategy adaptive\n\n# Monitor load distribution\nnpx claude-flow agent-metrics --type load-balancer\n\n# Adjust balancing parameters\nnpx claude-flow config-manage --action update --config '{\"stealThreshold\": 5, \"agingBoost\": 10}'\n```\n\n### Performance Monitoring\n```bash\n# Real-time load monitoring\nnpx claude-flow performance-report --format detailed\n\n# Bottleneck analysis\nnpx claude-flow bottleneck-analyze --component swarm-coordination\n\n# Resource utilization tracking\nnpx claude-flow metrics-collect --components [\"load-balancer\", \"task-queue\"]\n```\n\n## Integration Points\n\n### With Other Optimization Agents\n- **Performance Monitor**: Provides real-time metrics for load balancing decisions\n- **Topology Optimizer**: Coordinates topology changes based on load patterns\n- **Resource Allocator**: Optimizes resource distribution across the swarm\n\n### With Swarm Infrastructure\n- **Task Orchestrator**: Receives load-balanced task assignments\n- **Agent Coordinator**: Provides agent capacity and availability information\n- **Memory System**: Stores load balancing history and patterns\n\n## Performance Metrics\n\n### Key Performance Indicators\n- **Load Distribution Variance**: Measure of load balance across agents\n- **Task Migration Rate**: Frequency of work-stealing operations\n- **Queue Latency**: Average time tasks spend in queues\n- **Utilization Efficiency**: Percentage of optimal resource utilization\n- **Fairness Index**: Measure of fair resource allocation\n\n### Benchmarking\n```javascript\n\u002F\u002F Load balancer benchmarking suite\nconst benchmarks = {\n  async throughputTest(taskCount, agentCount) {\n    const startTime = performance.now();\n    await this.distributeAndExecute(taskCount, agentCount);\n    const endTime = performance.now();\n    \n    return {\n      throughput: taskCount \u002F ((endTime - startTime) \u002F 1000),\n      averageLatency: (endTime - startTime) \u002F taskCount\n    };\n  },\n  \n  async loadBalanceEfficiency(tasks, agents) {\n    const distribution = await this.distributeLoad(tasks, agents);\n    const idealLoad = tasks.length \u002F agents.length;\n    \n    const variance = distribution.reduce((sum, load) => \n      sum + Math.pow(load - idealLoad, 2), 0) \u002F agents.length;\n    \n    return {\n      efficiency: 1 \u002F (1 + variance),\n      loadVariance: variance\n    };\n  }\n};\n```\n\nThis Load Balancing Coordinator agent provides comprehensive task distribution optimization with advanced algorithms, real-time monitoring, and adaptive resource allocation capabilities for high-performance swarm coordination.",{"name":5,"description":7},"main","d69096e239f2a89aabc2cd078c47664a9d3b1e11","05-技能与代理开发\u002Fagent-load-balancer",3,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts","https:\u002F\u002Fgithub.com\u002Flza6","个人vx：Tf00798\r\n个人✈：@TFGY999","2026-10-08 14:29:22",{"source":8,"skillName":5,"commands":33,"prompt":43,"zipAvailable":44},[34,37,40],{"tool":35,"command":36},"npx","npx skills add lza6\u002FArmor-piercing-system-prompts --skill agent-load-balancer",{"tool":38,"command":39},"bunx","bunx skills add lza6\u002FArmor-piercing-system-prompts --skill agent-load-balancer",{"tool":41,"command":42},"pnpm","pnpm dlx skills add lza6\u002FArmor-piercing-system-prompts --skill agent-load-balancer","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F05-技能与代理开发\u002Fagent-load-balancer\u002FSKILL.md",true,[],[47,55,60,67,74,81],{"skillNo":48,"name":49,"nameZh":6,"description":50,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":51,"categories":52,"keywords":53,"publishedAt":54},"01M7FFG485KY83TD6HSC4YPCMC","obsidian-clipper-template-creator","Guide for creating templates for the Obsidian Web Clipper. Use when you want to create a new clipping template, understand available variables, or format clipped content.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator\u002FSKILL.md",[],[],"2026-11-16 10:05:34",{"skillNo":56,"name":49,"nameZh":6,"description":50,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":57,"categories":58,"keywords":59,"publishedAt":54},"01M7FFG0T6AZ0STEJPSB3CX84K","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator-antigravity-awesome-skills-main\u002FSKILL.md",[],[],{"skillNo":61,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":64,"categories":65,"keywords":66,"publishedAt":54},"01M7FFFY0QEM14SKWPDYGMXM2P","python-api","Programmatic Object-Detection-Metrics API guidance for BoundingBox, BoundingBoxes, Evaluator, enums, coordinate conversion, AP\u002FmAP semantics, and source-style import diagnostics.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Fpython-api\u002FSKILL.md",[],[],{"skillNo":68,"name":69,"nameZh":6,"description":70,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":71,"categories":72,"keywords":73,"publishedAt":54},"01M7FFFVZQZ2GJ0C0TGP7JHD20","file-evaluation","Folder-based PASCAL VOC AP\u002FmAP evaluation from ground-truth and detection text files.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Ffile-evaluation\u002FSKILL.md",[],[],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":54},"01M7FFFSFNYSE77SJ6PE1CK2DA","object-detection-metrics","Guides legacy Object-Detection-Metrics PASCAL VOC AP\u002FmAP evaluation from detection text folders and direct Python API classes.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":54},"01M7FFFP0GCK1RVXBCPQ8APBMT","nx-workspace-patterns","Configure and optimize Nx monorepo workspaces. Use when setting up Nx, configuring project boundaries, optimizing build caching, or implementing affected commands.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fnx-workspace-patterns\u002FSKILL.md",[],[]]