@claude-flow/swarm 3.0.0-alpha.1

package/src/queen-coordinator.ts

@@ -0,0 +1,2024 @@
+/**
+ * Queen Coordinator - Central Orchestrator for 15-Agent Swarm
+ *
+ * The Queen Coordinator is the strategic decision-maker for the V3 hive-mind system.
+ * It analyzes tasks, delegates to appropriate agents, monitors swarm health,
+ * coordinates consensus, and learns from outcomes using ReasoningBank patterns.
+ *
+ * Features:
+ * - Strategic task analysis with ReasoningBank pattern matching
+ * - Agent delegation with capability scoring and load balancing
+ * - Swarm health monitoring with bottleneck detection
+ * - Consensus coordination (majority, weighted, unanimous)
+ * - Learning from outcomes for continuous improvement
+ *
+ * Performance Targets:
+ * - Task analysis: <50ms
+ * - Agent scoring: <20ms
+ * - Consensus coordination: <100ms
+ * - Health check: <30ms
+ *
+ * @module @claude-flow/swarm/queen-coordinator
+ */
+
+import { EventEmitter } from 'events';
+import type {
+  AgentState,
+  AgentType,
+  AgentId,
+  TaskDefinition,
+  TaskType,
+  TaskPriority,
+  TaskStatus,
+  CoordinatorMetrics,
+  ConsensusResult,
+  SwarmEvent,
+  SwarmEventType,
+} from './types.js';
+import type { AgentDomain, DomainConfig, DomainStatus } from './unified-coordinator.js';
+
+// =============================================================================
+// Types
+// =============================================================================
+
+/**
+ * Task analysis result from the Queen
+ */
+export interface TaskAnalysis {
+  /** Unique analysis ID */
+  analysisId: string;
+  /** Original task ID */
+  taskId: string;
+  /** Task complexity score (0-1) */
+  complexity: number;
+  /** Estimated duration in milliseconds */
+  estimatedDurationMs: number;
+  /** Required capabilities for this task */
+  requiredCapabilities: string[];
+  /** Recommended domain for execution */
+  recommendedDomain: AgentDomain;
+  /** Sub-tasks if decomposition is needed */
+  subtasks: SubTask[];
+  /** Patterns found from ReasoningBank */
+  matchedPatterns: MatchedPattern[];
+  /** Resource requirements */
+  resourceRequirements: ResourceRequirements;
+  /** Confidence in this analysis (0-1) */
+  confidence: number;
+  /** Analysis timestamp */
+  timestamp: Date;
+}
+
+/**
+ * Sub-task from task decomposition
+ */
+export interface SubTask {
+  id: string;
+  name: string;
+  description: string;
+  type: TaskType;
+  priority: TaskPriority;
+  dependencies: string[];
+  estimatedDurationMs: number;
+  requiredCapabilities: string[];
+  recommendedDomain: AgentDomain;
+}
+
+/**
+ * Pattern matched from ReasoningBank
+ */
+export interface MatchedPattern {
+  patternId: string;
+  strategy: string;
+  successRate: number;
+  relevanceScore: number;
+  keyLearnings: string[];
+}
+
+/**
+ * Resource requirements for a task
+ */
+export interface ResourceRequirements {
+  minAgents: number;
+  maxAgents: number;
+  memoryMb: number;
+  cpuIntensive: boolean;
+  ioIntensive: boolean;
+  networkRequired: boolean;
+}
+
+/**
+ * Delegation plan for task execution
+ */
+export interface DelegationPlan {
+  /** Plan ID */
+  planId: string;
+  /** Task ID being delegated */
+  taskId: string;
+  /** Analysis that informed this plan */
+  analysisId: string;
+  /** Primary agent assignment */
+  primaryAgent: AgentAssignment;
+  /** Backup agents for failover */
+  backupAgents: AgentAssignment[];
+  /** Parallel sub-task assignments */
+  parallelAssignments: ParallelAssignment[];
+  /** Execution strategy */
+  strategy: ExecutionStrategy;
+  /** Estimated completion time */
+  estimatedCompletionMs: number;
+  /** Plan creation timestamp */
+  timestamp: Date;
+}
+
+/**
+ * Agent assignment in a delegation plan
+ */
+export interface AgentAssignment {
+  agentId: string;
+  domain: AgentDomain;
+  taskId: string;
+  score: number;
+  assignedAt: Date;
+}
+
+/**
+ * Parallel task assignment
+ */
+export interface ParallelAssignment {
+  subtaskId: string;
+  agentId: string;
+  domain: AgentDomain;
+  dependencies: string[];
+}
+
+/**
+ * Execution strategy for delegation
+ */
+export type ExecutionStrategy =
+  | 'sequential'
+  | 'parallel'
+  | 'pipeline'
+  | 'fan-out-fan-in'
+  | 'hybrid';
+
+/**
+ * Agent score for task assignment
+ */
+export interface AgentScore {
+  agentId: string;
+  domain: AgentDomain;
+  totalScore: number;
+  capabilityScore: number;
+  loadScore: number;
+  performanceScore: number;
+  healthScore: number;
+  availabilityScore: number;
+}
+
+/**
+ * Health report for the swarm
+ */
+export interface HealthReport {
+  /** Report ID */
+  reportId: string;
+  /** Report timestamp */
+  timestamp: Date;
+  /** Overall swarm health (0-1) */
+  overallHealth: number;
+  /** Status of each domain */
+  domainHealth: Map<AgentDomain, DomainHealthStatus>;
+  /** Individual agent health */
+  agentHealth: AgentHealthEntry[];
+  /** Detected bottlenecks */
+  bottlenecks: Bottleneck[];
+  /** Active alerts */
+  alerts: HealthAlert[];
+  /** Performance metrics */
+  metrics: HealthMetrics;
+  /** Recommendations for improvement */
+  recommendations: string[];
+}
+
+/**
+ * Domain health status
+ */
+export interface DomainHealthStatus {
+  domain: AgentDomain;
+  health: number;
+  activeAgents: number;
+  totalAgents: number;
+  queuedTasks: number;
+  avgResponseTimeMs: number;
+  errorRate: number;
+}
+
+/**
+ * Agent health entry
+ */
+export interface AgentHealthEntry {
+  agentId: string;
+  domain: AgentDomain;
+  health: number;
+  status: string;
+  lastHeartbeat: Date;
+  currentLoad: number;
+  recentErrors: number;
+}
+
+/**
+ * Bottleneck detection result
+ */
+export interface Bottleneck {
+  type: 'agent' | 'domain' | 'task' | 'resource';
+  location: string;
+  severity: 'low' | 'medium' | 'high' | 'critical';
+  description: string;
+  impact: string;
+  suggestedAction: string;
+}
+
+/**
+ * Health alert
+ */
+export interface HealthAlert {
+  alertId: string;
+  type: 'warning' | 'error' | 'critical';
+  source: string;
+  message: string;
+  timestamp: Date;
+  acknowledged: boolean;
+}
+
+/**
+ * Health metrics
+ */
+export interface HealthMetrics {
+  totalAgents: number;
+  activeAgents: number;
+  idleAgents: number;
+  errorAgents: number;
+  totalTasks: number;
+  completedTasks: number;
+  failedTasks: number;
+  avgTaskDurationMs: number;
+  taskThroughputPerMin: number;
+  consensusSuccessRate: number;
+}
+
+/**
+ * Decision requiring consensus
+ */
+export interface Decision {
+  decisionId: string;
+  type: DecisionType;
+  proposal: unknown;
+  requiredConsensus: ConsensusType;
+  timeout: number;
+  initiator: string;
+  metadata: Record<string, unknown>;
+}
+
+/**
+ * Decision types
+ */
+export type DecisionType =
+  | 'task-assignment'
+  | 'resource-allocation'
+  | 'topology-change'
+  | 'agent-termination'
+  | 'priority-override'
+  | 'emergency-action';
+
+/**
+ * Consensus types
+ */
+export type ConsensusType =
+  | 'majority'
+  | 'supermajority'
+  | 'unanimous'
+  | 'weighted'
+  | 'queen-override';
+
+/**
+ * Task result for learning
+ */
+export interface TaskResult {
+  taskId: string;
+  success: boolean;
+  output?: unknown;
+  error?: string;
+  durationMs: number;
+  agentId: string;
+  domain: AgentDomain;
+  metrics: TaskMetrics;
+}
+
+/**
+ * Task execution metrics
+ */
+export interface TaskMetrics {
+  startTime: Date;
+  endTime: Date;
+  retries: number;
+  resourceUsage: {
+    memoryMb: number;
+    cpuPercent: number;
+  };
+  stepsCompleted: number;
+  qualityScore: number;
+}
+
+/**
+ * Queen Coordinator configuration
+ */
+export interface QueenCoordinatorConfig {
+  /** Enable ReasoningBank integration */
+  enableLearning: boolean;
+  /** Number of patterns to retrieve for analysis */
+  patternRetrievalK: number;
+  /** Minimum pattern relevance threshold */
+  patternThreshold: number;
+  /** Task complexity thresholds */
+  complexityThresholds: {
+    simple: number;
+    moderate: number;
+    complex: number;
+  };
+  /** Health check interval in ms */
+  healthCheckIntervalMs: number;
+  /** Bottleneck detection thresholds */
+  bottleneckThresholds: {
+    queueDepth: number;
+    errorRate: number;
+    responseTimeMs: number;
+  };
+  /** Consensus timeouts */
+  consensusTimeouts: {
+    majority: number;
+    supermajority: number;
+    unanimous: number;
+  };
+  /** Enable automatic failover */
+  enableFailover: boolean;
+  /** Maximum delegation attempts */
+  maxDelegationAttempts: number;
+}
+
+/**
+ * Default Queen Coordinator configuration
+ */
+const DEFAULT_CONFIG: QueenCoordinatorConfig = {
+  enableLearning: true,
+  patternRetrievalK: 5,
+  patternThreshold: 0.6,
+  complexityThresholds: {
+    simple: 0.3,
+    moderate: 0.6,
+    complex: 0.85,
+  },
+  healthCheckIntervalMs: 10000,
+  bottleneckThresholds: {
+    queueDepth: 10,
+    errorRate: 0.1,
+    responseTimeMs: 5000,
+  },
+  consensusTimeouts: {
+    majority: 5000,
+    supermajority: 10000,
+    unanimous: 30000,
+  },
+  enableFailover: true,
+  maxDelegationAttempts: 3,
+};
+
+// =============================================================================
+// Interfaces for Dependencies
+// =============================================================================
+
+/**
+ * Interface for swarm coordinator interactions
+ */
+export interface ISwarmCoordinator {
+  getAgentsByDomain(domain: AgentDomain): AgentState[];
+  getAllAgents(): AgentState[];
+  getAvailableAgents(): AgentState[];
+  getMetrics(): CoordinatorMetrics;
+  getDomainConfigs(): Map<AgentDomain, DomainConfig>;
+  getStatus(): {
+    domains: DomainStatus[];
+    metrics: CoordinatorMetrics;
+  };
+  assignTaskToDomain(taskId: string, domain: AgentDomain): Promise<string | undefined>;
+  proposeConsensus(value: unknown): Promise<ConsensusResult>;
+  broadcastMessage(payload: unknown, priority?: 'urgent' | 'high' | 'normal' | 'low'): Promise<void>;
+}
+
+/**
+ * Interface for neural learning system interactions
+ */
+export interface INeuralLearningSystem {
+  initialize(): Promise<void>;
+  beginTask(context: string, domain?: string): string;
+  recordStep(trajectoryId: string, action: string, reward: number, stateEmbedding: Float32Array): void;
+  completeTask(trajectoryId: string, quality?: number): Promise<void>;
+  findPatterns(queryEmbedding: Float32Array, k?: number): Promise<PatternMatchResult[]>;
+  retrieveMemories(queryEmbedding: Float32Array, k?: number): Promise<MemoryRetrievalResult[]>;
+  triggerLearning(): Promise<void>;
+}
+
+/**
+ * Pattern match result from neural system
+ */
+export interface PatternMatchResult {
+  patternId: string;
+  strategy: string;
+  successRate: number;
+  relevanceScore: number;
+  keyLearnings?: string[];
+}
+
+/**
+ * Memory retrieval result from neural system
+ */
+export interface MemoryRetrievalResult {
+  memory: {
+    memoryId: string;
+    strategy: string;
+    quality: number;
+    keyLearnings: string[];
+  };
+  relevanceScore: number;
+  combinedScore: number;
+}
+
+/**
+ * Interface for memory service interactions
+ */
+export interface IMemoryService {
+  semanticSearch(query: string, k?: number): Promise<SearchResultEntry[]>;
+  store(entry: MemoryStoreEntry): Promise<void>;
+}
+
+/**
+ * Search result entry from memory service
+ */
+export interface SearchResultEntry {
+  entry: {
+    id: string;
+    content: string;
+    metadata: Record<string, unknown>;
+  };
+  score: number;
+}
+
+/**
+ * Memory store entry
+ */
+export interface MemoryStoreEntry {
+  key: string;
+  content: string;
+  namespace: string;
+  tags: string[];
+  metadata: Record<string, unknown>;
+}
+
+// =============================================================================
+// Queen Coordinator Class
+// =============================================================================
+
+/**
+ * Queen Coordinator - Central orchestrator for the 15-agent hive-mind swarm
+ *
+ * The Queen is responsible for:
+ * 1. Strategic task analysis and decomposition
+ * 2. Agent delegation with load balancing
+ * 3. Swarm health monitoring
+ * 4. Consensus coordination
+ * 5. Learning from outcomes
+ */
+export class QueenCoordinator extends EventEmitter {
+  private config: QueenCoordinatorConfig;
+  private swarm: ISwarmCoordinator;
+  private neural?: INeuralLearningSystem;
+  private memory?: IMemoryService;
+
+  // Internal state
+  private analysisCache: Map<string, TaskAnalysis> = new Map();
+  private delegationPlans: Map<string, DelegationPlan> = new Map();
+  private activeDecisions: Map<string, Decision> = new Map();
+  private outcomeHistory: TaskResult[] = [];
+  private healthHistory: HealthReport[] = [];
+
+  // Counters for IDs
+  private analysisCounter = 0;
+  private planCounter = 0;
+  private reportCounter = 0;
+  private decisionCounter = 0;
+
+  // Health monitoring
+  private healthCheckInterval?: NodeJS.Timeout;
+  private lastHealthReport?: HealthReport;
+
+  // Performance tracking
+  private analysisLatencies: number[] = [];
+  private delegationLatencies: number[] = [];
+  private consensusLatencies: number[] = [];
+
+  constructor(
+    swarm: ISwarmCoordinator,
+    config: Partial<QueenCoordinatorConfig> = {},
+    neural?: INeuralLearningSystem,
+    memory?: IMemoryService
+  ) {
+    super();
+    this.swarm = swarm;
+    this.config = { ...DEFAULT_CONFIG, ...config };
+    this.neural = neural;
+    this.memory = memory;
+  }
+
+  // ===========================================================================
+  // Lifecycle
+  // ===========================================================================
+
+  /**
+   * Initialize the Queen Coordinator
+   */
+  async initialize(): Promise<void> {
+    // Initialize neural system if available
+    if (this.neural && this.config.enableLearning) {
+      await this.neural.initialize();
+    }
+
+    // Start health monitoring
+    this.startHealthMonitoring();
+
+    this.emitEvent('queen.initialized', {
+      config: this.config,
+      learningEnabled: this.config.enableLearning && !!this.neural,
+    });
+  }
+
+  /**
+   * Shutdown the Queen Coordinator
+   */
+  async shutdown(): Promise<void> {
+    this.stopHealthMonitoring();
+
+    // Trigger final learning if enabled
+    if (this.neural && this.config.enableLearning) {
+      await this.neural.triggerLearning();
+    }
+
+    this.emitEvent('queen.shutdown', {
+      totalAnalyses: this.analysisCounter,
+      totalPlans: this.planCounter,
+      totalDecisions: this.decisionCounter,
+    });
+  }
+
+  // ===========================================================================
+  // Strategic Task Analysis
+  // ===========================================================================
+
+  /**
+   * Analyze a task for optimal execution
+   *
+   * @param task - Task to analyze
+   * @returns Task analysis with recommendations
+   */
+  async analyzeTask(task: TaskDefinition): Promise<TaskAnalysis> {
+    const startTime = performance.now();
+
+    this.analysisCounter++;
+    const analysisId = `analysis_${Date.now()}_${this.analysisCounter}`;
+
+    // Decompose complex tasks
+    const subtasks = this.decomposeTask(task);
+
+    // Identify required capabilities
+    const requiredCapabilities = this.identifyRequiredCapabilities(task);
+
+    // Calculate complexity
+    const complexity = this.calculateComplexity(task, subtasks);
+
+    // Estimate duration
+    const estimatedDurationMs = this.estimateDuration(task, complexity, subtasks);
+
+    // Determine recommended domain
+    const recommendedDomain = this.determineOptimalDomain(task, requiredCapabilities);
+
+    // Find matching patterns from ReasoningBank
+    const matchedPatterns = await this.findMatchingPatterns(task);
+
+    // Estimate resource requirements
+    const resourceRequirements = this.estimateResources(task, complexity);
+
+    // Calculate confidence
+    const confidence = this.calculateAnalysisConfidence(
+      matchedPatterns,
+      complexity,
+      requiredCapabilities
+    );
+
+    const analysis: TaskAnalysis = {
+      analysisId,
+      taskId: task.id.id,
+      complexity,
+      estimatedDurationMs,
+      requiredCapabilities,
+      recommendedDomain,
+      subtasks,
+      matchedPatterns,
+      resourceRequirements,
+      confidence,
+      timestamp: new Date(),
+    };
+
+    // Cache the analysis
+    this.analysisCache.set(analysisId, analysis);
+
+    // Record latency
+    const latency = performance.now() - startTime;
+    this.analysisLatencies.push(latency);
+    if (this.analysisLatencies.length > 100) {
+      this.analysisLatencies.shift();
+    }
+
+    this.emitEvent('queen.task.analyzed', {
+      analysisId,
+      taskId: task.id.id,
+      complexity,
+      recommendedDomain,
+      patternsFound: matchedPatterns.length,
+      latencyMs: latency,
+    });
+
+    return analysis;
+  }
+
+  /**
+   * Decompose a complex task into subtasks
+   */
+  private decomposeTask(task: TaskDefinition): SubTask[] {
+    const subtasks: SubTask[] = [];
+
+    // Simple tasks don't need decomposition
+    if (this.isSimpleTask(task)) {
+      return subtasks;
+    }
+
+    // Decompose based on task type
+    switch (task.type) {
+      case 'coding':
+        subtasks.push(...this.decomposeCodingTask(task));
+        break;
+      case 'testing':
+        subtasks.push(...this.decomposeTestingTask(task));
+        break;
+      case 'research':
+        subtasks.push(...this.decomposeResearchTask(task));
+        break;
+      case 'coordination':
+        subtasks.push(...this.decomposeCoordinationTask(task));
+        break;
+      default:
+        // Generic decomposition
+        subtasks.push(...this.decomposeGenericTask(task));
+    }
+
+    return subtasks;
+  }
+
+  private isSimpleTask(task: TaskDefinition): boolean {
+    // Estimate if task is simple based on description length and type
+    const descLength = task.description?.length || 0;
+    const isSimpleType = ['documentation', 'review'].includes(task.type);
+    return descLength < 200 || isSimpleType;
+  }
+
+  private decomposeCodingTask(task: TaskDefinition): SubTask[] {
+    return [
+      {
+        id: `${task.id.id}_design`,
+        name: 'Design & Planning',
+        description: 'Design the solution architecture',
+        type: 'analysis',
+        priority: task.priority,
+        dependencies: [],
+        estimatedDurationMs: 10000,
+        requiredCapabilities: ['design', 'architecture'],
+        recommendedDomain: 'core',
+      },
+      {
+        id: `${task.id.id}_implement`,
+        name: 'Implementation',
+        description: 'Implement the designed solution',
+        type: 'coding',
+        priority: task.priority,
+        dependencies: [`${task.id.id}_design`],
+        estimatedDurationMs: 30000,
+        requiredCapabilities: ['coding', 'implementation'],
+        recommendedDomain: 'integration',
+      },
+      {
+        id: `${task.id.id}_test`,
+        name: 'Testing',
+        description: 'Test the implementation',
+        type: 'testing',
+        priority: task.priority,
+        dependencies: [`${task.id.id}_implement`],
+        estimatedDurationMs: 15000,
+        requiredCapabilities: ['testing', 'validation'],
+        recommendedDomain: 'support',
+      },
+    ];
+  }
+
+  private decomposeTestingTask(task: TaskDefinition): SubTask[] {
+    return [
+      {
+        id: `${task.id.id}_analyze`,
+        name: 'Test Analysis',
+        description: 'Analyze what needs to be tested',
+        type: 'analysis',
+        priority: task.priority,
+        dependencies: [],
+        estimatedDurationMs: 5000,
+        requiredCapabilities: ['analysis', 'testing'],
+        recommendedDomain: 'support',
+      },
+      {
+        id: `${task.id.id}_execute`,
+        name: 'Test Execution',
+        description: 'Execute the tests',
+        type: 'testing',
+        priority: task.priority,
+        dependencies: [`${task.id.id}_analyze`],
+        estimatedDurationMs: 20000,
+        requiredCapabilities: ['testing', 'execution'],
+        recommendedDomain: 'support',
+      },
+    ];
+  }
+
+  private decomposeResearchTask(task: TaskDefinition): SubTask[] {
+    return [
+      {
+        id: `${task.id.id}_gather`,
+        name: 'Information Gathering',
+        description: 'Gather relevant information',
+        type: 'research',
+        priority: task.priority,
+        dependencies: [],
+        estimatedDurationMs: 15000,
+        requiredCapabilities: ['research', 'analysis'],
+        recommendedDomain: 'core',
+      },
+      {
+        id: `${task.id.id}_analyze`,
+        name: 'Analysis',
+        description: 'Analyze gathered information',
+        type: 'analysis',
+        priority: task.priority,
+        dependencies: [`${task.id.id}_gather`],
+        estimatedDurationMs: 10000,
+        requiredCapabilities: ['analysis', 'synthesis'],
+        recommendedDomain: 'core',
+      },
+    ];
+  }
+
+  private decomposeCoordinationTask(task: TaskDefinition): SubTask[] {
+    return [
+      {
+        id: `${task.id.id}_plan`,
+        name: 'Planning',
+        description: 'Create coordination plan',
+        type: 'coordination',
+        priority: task.priority,
+        dependencies: [],
+        estimatedDurationMs: 5000,
+        requiredCapabilities: ['planning', 'coordination'],
+        recommendedDomain: 'queen',
+      },
+      {
+        id: `${task.id.id}_execute`,
+        name: 'Execution',
+        description: 'Execute coordination plan',
+        type: 'coordination',
+        priority: task.priority,
+        dependencies: [`${task.id.id}_plan`],
+        estimatedDurationMs: 10000,
+        requiredCapabilities: ['coordination', 'oversight'],
+        recommendedDomain: 'queen',
+      },
+    ];
+  }
+
+  private decomposeGenericTask(task: TaskDefinition): SubTask[] {
+    return [
+      {
+        id: `${task.id.id}_execute`,
+        name: 'Task Execution',
+        description: task.description,
+        type: task.type,
+        priority: task.priority,
+        dependencies: [],
+        estimatedDurationMs: 20000,
+        requiredCapabilities: this.identifyRequiredCapabilities(task),
+        recommendedDomain: this.inferDomainFromType(task.type),
+      },
+    ];
+  }
+
+  /**
+   * Identify required capabilities for a task
+   */
+  private identifyRequiredCapabilities(task: TaskDefinition): string[] {
+    const capabilities: Set<string> = new Set();
+
+    // Add type-based capabilities
+    const typeCapabilities: Record<TaskType, string[]> = {
+      research: ['research', 'analysis', 'synthesis'],
+      analysis: ['analysis', 'reasoning', 'synthesis'],
+      coding: ['coding', 'implementation', 'debugging'],
+      testing: ['testing', 'validation', 'quality'],
+      review: ['review', 'analysis', 'feedback'],
+      documentation: ['documentation', 'writing', 'clarity'],
+      coordination: ['coordination', 'planning', 'oversight'],
+      consensus: ['consensus', 'voting', 'agreement'],
+      custom: ['general', 'execution'],
+    };
+
+    for (const cap of typeCapabilities[task.type] || ['general']) {
+      capabilities.add(cap);
+    }
+
+    // Extract additional capabilities from description
+    const descLower = (task.description || '').toLowerCase();
+    if (descLower.includes('security')) capabilities.add('security');
+    if (descLower.includes('performance')) capabilities.add('performance');
+    if (descLower.includes('architecture')) capabilities.add('architecture');
+    if (descLower.includes('integration')) capabilities.add('integration');
+    if (descLower.includes('deploy')) capabilities.add('deployment');
+
+    return Array.from(capabilities);
+  }
+
+  /**
+   * Calculate task complexity score
+   */
+  private calculateComplexity(task: TaskDefinition, subtasks: SubTask[]): number {
+    let complexity = 0.3; // Base complexity
+
+    // Add complexity for subtasks
+    complexity += subtasks.length * 0.1;
+
+    // Add complexity for dependencies
+    const totalDeps = subtasks.reduce((sum, st) => sum + st.dependencies.length, 0);
+    complexity += totalDeps * 0.05;
+
+    // Add complexity for priority
+    const priorityMultipliers: Record<TaskPriority, number> = {
+      critical: 1.3,
+      high: 1.15,
+      normal: 1.0,
+      low: 0.9,
+      background: 0.8,
+    };
+    complexity *= priorityMultipliers[task.priority];
+
+    // Add complexity for task type
+    const typeComplexity: Record<TaskType, number> = {
+      coordination: 0.2,
+      consensus: 0.25,
+      coding: 0.15,
+      testing: 0.1,
+      analysis: 0.1,
+      research: 0.1,
+      review: 0.05,
+      documentation: 0.05,
+      custom: 0.1,
+    };
+    complexity += typeComplexity[task.type] || 0.1;
+
+    // Add complexity for description length
+    const descLength = task.description?.length || 0;
+    complexity += Math.min(descLength / 2000, 0.2);
+
+    return Math.min(complexity, 1.0);
+  }
+
+  /**
+   * Estimate task duration
+   */
+  private estimateDuration(
+    task: TaskDefinition,
+    complexity: number,
+    subtasks: SubTask[]
+  ): number {
+    // Base duration by type (in ms)
+    const baseDurations: Record<TaskType, number> = {
+      research: 30000,
+      analysis: 20000,
+      coding: 60000,
+      testing: 30000,
+      review: 15000,
+      documentation: 20000,
+      coordination: 10000,
+      consensus: 15000,
+      custom: 30000,
+    };
+
+    const baseDuration = baseDurations[task.type] || 30000;
+
+    // Adjust for complexity
+    const complexityMultiplier = 0.5 + complexity * 1.5;
+
+    // Add subtask durations
+    const subtaskDuration = subtasks.reduce((sum, st) => sum + st.estimatedDurationMs, 0);
+
+    // Total estimate
+    return Math.round(baseDuration * complexityMultiplier + subtaskDuration * 0.5);
+  }
+
+  /**
+   * Determine optimal domain for task execution
+   */
+  private determineOptimalDomain(
+    task: TaskDefinition,
+    capabilities: string[]
+  ): AgentDomain {
+    // Check capabilities for domain hints
+    if (capabilities.includes('security')) return 'security';
+    if (capabilities.includes('coordination') || capabilities.includes('planning')) return 'queen';
+    if (capabilities.includes('testing') || capabilities.includes('performance')) return 'support';
+    if (capabilities.includes('integration')) return 'integration';
+
+    // Fall back to type-based inference
+    return this.inferDomainFromType(task.type);
+  }
+
+  private inferDomainFromType(type: TaskType): AgentDomain {
+    const typeDomains: Record<TaskType, AgentDomain> = {
+      coordination: 'queen',
+      consensus: 'queen',
+      coding: 'integration',
+      testing: 'support',
+      review: 'security',
+      analysis: 'core',
+      research: 'core',
+      documentation: 'support',
+      custom: 'core',
+    };
+    return typeDomains[type] || 'core';
+  }
+
+  /**
+   * Find matching patterns from ReasoningBank
+   */
+  private async findMatchingPatterns(task: TaskDefinition): Promise<MatchedPattern[]> {
+    const patterns: MatchedPattern[] = [];
+
+    if (!this.neural || !this.config.enableLearning) {
+      return patterns;
+    }
+
+    try {
+      // Create a simple embedding from task description
+      const embedding = this.createSimpleEmbedding(task.description || task.name);
+
+      // Query ReasoningBank for similar patterns
+      const results = await this.neural.findPatterns(embedding, this.config.patternRetrievalK);
+
+      for (const result of results) {
+        if (result.relevanceScore >= this.config.patternThreshold) {
+          patterns.push({
+            patternId: result.patternId,
+            strategy: result.strategy,
+            successRate: result.successRate,
+            relevanceScore: result.relevanceScore,
+            keyLearnings: result.keyLearnings || [],
+          });
+        }
+      }
+    } catch (error) {
+      // Log but don't fail - patterns are optional
+      this.emitEvent('queen.pattern.error', { error: String(error) });
+    }
+
+    return patterns;
+  }
+
+  /**
+   * Create a simple embedding from text (placeholder)
+   */
+  private createSimpleEmbedding(text: string): Float32Array {
+    // Simple hash-based embedding for demonstration
+    // In production, this would use a real embedding model
+    const embedding = new Float32Array(768);
+    const words = text.toLowerCase().split(/\s+/);
+
+    for (let i = 0; i < words.length; i++) {
+      const word = words[i];
+      for (let j = 0; j < word.length; j++) {
+        const idx = (word.charCodeAt(j) * (i + 1) * (j + 1)) % 768;
+        embedding[idx] += 1 / words.length;
+      }
+    }
+
+    // Normalize
+    let norm = 0;
+    for (let i = 0; i < embedding.length; i++) {
+      norm += embedding[i] * embedding[i];
+    }
+    norm = Math.sqrt(norm);
+    if (norm > 0) {
+      for (let i = 0; i < embedding.length; i++) {
+        embedding[i] /= norm;
+      }
+    }
+
+    return embedding;
+  }
+
+  /**
+   * Estimate resource requirements
+   */
+  private estimateResources(task: TaskDefinition, complexity: number): ResourceRequirements {
+    const minAgents = complexity > 0.7 ? 2 : 1;
+    const maxAgents = complexity > 0.8 ? 4 : complexity > 0.5 ? 3 : 2;
+
+    return {
+      minAgents,
+      maxAgents,
+      memoryMb: Math.round(256 + complexity * 512),
+      cpuIntensive: ['coding', 'analysis'].includes(task.type),
+      ioIntensive: ['research', 'testing'].includes(task.type),
+      networkRequired: task.type === 'research',
+    };
+  }
+
+  /**
+   * Calculate confidence in analysis
+   */
+  private calculateAnalysisConfidence(
+    patterns: MatchedPattern[],
+    complexity: number,
+    capabilities: string[]
+  ): number {
+    let confidence = 0.5; // Base confidence
+
+    // Higher confidence with more matching patterns
+    confidence += patterns.length * 0.1;
+
+    // Higher confidence for simpler tasks
+    confidence += (1 - complexity) * 0.2;
+
+    // Higher confidence with clear capabilities
+    confidence += Math.min(capabilities.length * 0.05, 0.15);
+
+    // Boost from high-success patterns
+    const avgPatternSuccess = patterns.length > 0
+      ? patterns.reduce((sum, p) => sum + p.successRate, 0) / patterns.length
+      : 0;
+    confidence += avgPatternSuccess * 0.1;
+
+    return Math.min(confidence, 0.95);
+  }
+
+  // ===========================================================================
+  // Agent Delegation
+  // ===========================================================================
+
+  /**
+   * Delegate a task to agents based on analysis
+   *
+   * @param task - Task to delegate
+   * @param analysis - Previous task analysis
+   * @returns Delegation plan
+   */
+  async delegateToAgents(task: TaskDefinition, analysis: TaskAnalysis): Promise<DelegationPlan> {
+    const startTime = performance.now();
+
+    this.planCounter++;
+    const planId = `plan_${Date.now()}_${this.planCounter}`;
+
+    // Score all available agents for this task
+    const agentScores = this.scoreAgents(task, analysis.matchedPatterns);
+
+    // Select primary agent
+    const primaryAgent = this.selectPrimaryAgent(task, agentScores, analysis);
+
+    // Select backup agents
+    const backupAgents = this.selectBackupAgents(task, agentScores, primaryAgent);
+
+    // Create parallel assignments for subtasks
+    const parallelAssignments = this.createParallelAssignments(
+      analysis.subtasks,
+      agentScores
+    );
+
+    // Determine execution strategy
+    const strategy = this.determineExecutionStrategy(analysis);
+
+    const plan: DelegationPlan = {
+      planId,
+      taskId: task.id.id,
+      analysisId: analysis.analysisId,
+      primaryAgent,
+      backupAgents,
+      parallelAssignments,
+      strategy,
+      estimatedCompletionMs: analysis.estimatedDurationMs,
+      timestamp: new Date(),
+    };
+
+    // Store the plan
+    this.delegationPlans.set(planId, plan);
+
+    // Execute the delegation
+    await this.executeDelegation(plan);
+
+    // Record latency
+    const latency = performance.now() - startTime;
+    this.delegationLatencies.push(latency);
+    if (this.delegationLatencies.length > 100) {
+      this.delegationLatencies.shift();
+    }
+
+    this.emitEvent('queen.task.delegated', {
+      planId,
+      taskId: task.id.id,
+      primaryAgent: primaryAgent.agentId,
+      strategy,
+      latencyMs: latency,
+    });
+
+    return plan;
+  }
+
+  /**
+   * Score agents for task assignment
+   */
+  scoreAgents(task: TaskDefinition, patterns: MatchedPattern[]): AgentScore[] {
+    const agents = this.swarm.getAllAgents();
+    const scores: AgentScore[] = [];
+
+    for (const agent of agents) {
+      const score = this.scoreAgent(agent, task, patterns);
+      scores.push(score);
+    }
+
+    // Sort by total score descending
+    scores.sort((a, b) => b.totalScore - a.totalScore);
+
+    return scores;
+  }
+
+  private scoreAgent(
+    agent: AgentState,
+    task: TaskDefinition,
+    patterns: MatchedPattern[]
+  ): AgentScore {
+    const domain = this.getAgentDomain(agent);
+
+    // Capability score - how well agent capabilities match task
+    const capabilityScore = this.calculateCapabilityScore(agent, task);
+
+    // Load score - prefer less loaded agents
+    const loadScore = 1 - agent.workload;
+
+    // Performance score - based on past performance
+    const performanceScore = this.calculatePerformanceScore(agent, patterns);
+
+    // Health score - agent health
+    const healthScore = agent.health;
+
+    // Availability score - is agent actually available
+    const availabilityScore = agent.status === 'idle' ? 1.0 :
+                              agent.status === 'busy' ? 0.3 : 0.0;
+
+    // Weighted total
+    const totalScore =
+      capabilityScore * 0.30 +
+      loadScore * 0.20 +
+      performanceScore * 0.25 +
+      healthScore * 0.15 +
+      availabilityScore * 0.10;
+
+    return {
+      agentId: agent.id.id,
+      domain,
+      totalScore,
+      capabilityScore,
+      loadScore,
+      performanceScore,
+      healthScore,
+      availabilityScore,
+    };
+  }
+
+  private calculateCapabilityScore(agent: AgentState, task: TaskDefinition): number {
+    let score = 0.5; // Base score
+
+    // Check type match
+    const typeMatches: Record<TaskType, AgentType[]> = {
+      research: ['researcher'],
+      analysis: ['analyst', 'researcher'],
+      coding: ['coder'],
+      testing: ['tester'],
+      review: ['reviewer'],
+      documentation: ['documenter'],
+      coordination: ['coordinator', 'queen'],
+      consensus: ['coordinator', 'queen'],
+      custom: ['worker'],
+    };
+
+    const preferredTypes = typeMatches[task.type] || ['worker'];
+    if (preferredTypes.includes(agent.type)) {
+      score += 0.3;
+    }
+
+    // Check specific capabilities
+    const caps = agent.capabilities;
+    if (task.type === 'coding' && caps.codeGeneration) score += 0.1;
+    if (task.type === 'review' && caps.codeReview) score += 0.1;
+    if (task.type === 'testing' && caps.testing) score += 0.1;
+    if (task.type === 'coordination' && caps.coordination) score += 0.1;
+
+    return Math.min(score, 1.0);
+  }
+
+  private calculatePerformanceScore(
+    agent: AgentState,
+    patterns: MatchedPattern[]
+  ): number {
+    // Base on agent's success rate
+    let score = agent.metrics.successRate;
+
+    // Boost if patterns suggest this agent type is successful
+    const agentDomain = this.getAgentDomain(agent);
+    for (const pattern of patterns) {
+      if (pattern.strategy.toLowerCase().includes(agentDomain)) {
+        score += pattern.successRate * 0.1;
+      }
+    }
+
+    return Math.min(score, 1.0);
+  }
+
+  private getAgentDomain(agent: AgentState): AgentDomain {
+    // Determine domain from agent type
+    switch (agent.type) {
+      case 'queen':
+        return 'queen';
+      case 'specialist':
+        return 'security';
+      case 'architect':
+        return 'core';
+      case 'coder':
+        return 'integration';
+      case 'tester':
+        return 'support';
+      default:
+        return 'core';
+    }
+  }
+
+  private selectPrimaryAgent(
+    task: TaskDefinition,
+    scores: AgentScore[],
+    analysis: TaskAnalysis
+  ): AgentAssignment {
+    // Prefer agent from recommended domain
+    const domainAgents = scores.filter(s => s.domain === analysis.recommendedDomain);
+
+    // Fall back to best overall if no domain match
+    const bestScore = domainAgents.length > 0 ? domainAgents[0] : scores[0];
+
+    return {
+      agentId: bestScore.agentId,
+      domain: bestScore.domain,
+      taskId: task.id.id,
+      score: bestScore.totalScore,
+      assignedAt: new Date(),
+    };
+  }
+
+  private selectBackupAgents(
+    task: TaskDefinition,
+    scores: AgentScore[],
+    primary: AgentAssignment
+  ): AgentAssignment[] {
+    const backups: AgentAssignment[] = [];
+
+    // Select up to 2 backup agents
+    for (const score of scores) {
+      if (score.agentId === primary.agentId) continue;
+      if (score.totalScore < 0.3) continue;
+      if (backups.length >= 2) break;
+
+      backups.push({
+        agentId: score.agentId,
+        domain: score.domain,
+        taskId: task.id.id,
+        score: score.totalScore,
+        assignedAt: new Date(),
+      });
+    }
+
+    return backups;
+  }
+
+  private createParallelAssignments(
+    subtasks: SubTask[],
+    scores: AgentScore[]
+  ): ParallelAssignment[] {
+    const assignments: ParallelAssignment[] = [];
+
+    for (const subtask of subtasks) {
+      // Find best agent for this subtask's domain
+      const domainScores = scores.filter(s => s.domain === subtask.recommendedDomain);
+      const bestScore = domainScores.length > 0 ? domainScores[0] : scores[0];
+
+      assignments.push({
+        subtaskId: subtask.id,
+        agentId: bestScore.agentId,
+        domain: subtask.recommendedDomain,
+        dependencies: subtask.dependencies,
+      });
+    }
+
+    return assignments;
+  }
+
+  private determineExecutionStrategy(analysis: TaskAnalysis): ExecutionStrategy {
+    const subtaskCount = analysis.subtasks.length;
+
+    if (subtaskCount === 0) {
+      return 'sequential';
+    }
+
+    // Check for dependencies
+    const hasDependencies = analysis.subtasks.some(st => st.dependencies.length > 0);
+
+    if (!hasDependencies && subtaskCount > 2) {
+      return 'parallel';
+    }
+
+    if (hasDependencies && subtaskCount > 3) {
+      return 'pipeline';
+    }
+
+    if (analysis.complexity > 0.7) {
+      return 'fan-out-fan-in';
+    }
+
+    return 'hybrid';
+  }
+
+  private async executeDelegation(plan: DelegationPlan): Promise<void> {
+    // Delegate to the primary agent
+    await this.swarm.assignTaskToDomain(plan.taskId, plan.primaryAgent.domain);
+
+    // Notify about the delegation
+    await this.swarm.broadcastMessage(
+      {
+        type: 'delegation',
+        planId: plan.planId,
+        taskId: plan.taskId,
+        primaryAgent: plan.primaryAgent.agentId,
+        strategy: plan.strategy,
+      },
+      'normal'
+    );
+  }
+
+  // ===========================================================================
+  // Swarm Health Monitoring
+  // ===========================================================================
+
+  /**
+   * Monitor swarm health and detect issues
+   *
+   * @returns Health report
+   */
+  async monitorSwarmHealth(): Promise<HealthReport> {
+    this.reportCounter++;
+    const reportId = `health_${Date.now()}_${this.reportCounter}`;
+
+    const agents = this.swarm.getAllAgents();
+    const status = this.swarm.getStatus();
+    const metrics = this.swarm.getMetrics();
+
+    // Compute domain health
+    const domainHealth = this.computeDomainHealth(status.domains);
+
+    // Compute agent health
+    const agentHealth = this.computeAgentHealth(agents);
+
+    // Detect bottlenecks
+    const bottlenecks = this.detectBottlenecks(status, agents, metrics);
+
+    // Generate alerts
+    const alerts = this.generateAlerts(bottlenecks, agentHealth, metrics);
+
+    // Overall health score
+    const overallHealth = this.calculateOverallHealth(domainHealth, agentHealth, bottlenecks);
+
+    // Generate recommendations
+    const recommendations = this.generateRecommendations(bottlenecks, overallHealth);
+
+    // Compile health metrics
+    const healthMetrics: HealthMetrics = {
+      totalAgents: agents.length,
+      activeAgents: agents.filter(a => a.status === 'idle' || a.status === 'busy').length,
+      idleAgents: agents.filter(a => a.status === 'idle').length,
+      errorAgents: agents.filter(a => a.status === 'error').length,
+      totalTasks: metrics.totalTasks,
+      completedTasks: metrics.completedTasks,
+      failedTasks: metrics.failedTasks,
+      avgTaskDurationMs: metrics.avgTaskDurationMs,
+      taskThroughputPerMin: metrics.completedTasks / Math.max(metrics.uptime / 60, 1),
+      consensusSuccessRate: metrics.consensusSuccessRate,
+    };
+
+    const report: HealthReport = {
+      reportId,
+      timestamp: new Date(),
+      overallHealth,
+      domainHealth,
+      agentHealth,
+      bottlenecks,
+      alerts,
+      metrics: healthMetrics,
+      recommendations,
+    };
+
+    // Store in history
+    this.healthHistory.push(report);
+    if (this.healthHistory.length > 100) {
+      this.healthHistory.shift();
+    }
+    this.lastHealthReport = report;
+
+    this.emitEvent('queen.health.report', {
+      reportId,
+      overallHealth,
+      bottleneckCount: bottlenecks.length,
+      alertCount: alerts.length,
+    });
+
+    return report;
+  }
+
+  private computeDomainHealth(domains: DomainStatus[]): Map<AgentDomain, DomainHealthStatus> {
+    const health = new Map<AgentDomain, DomainHealthStatus>();
+
+    for (const domain of domains) {
+      const utilization = domain.agentCount > 0
+        ? domain.busyAgents / domain.agentCount
+        : 0;
+
+      const queuePressure = domain.tasksQueued > 5 ? 0.3 : domain.tasksQueued * 0.05;
+      const domainHealth = Math.max(0, 1 - queuePressure - (1 - utilization) * 0.2);
+
+      health.set(domain.name, {
+        domain: domain.name,
+        health: domainHealth,
+        activeAgents: domain.availableAgents + domain.busyAgents,
+        totalAgents: domain.agentCount,
+        queuedTasks: domain.tasksQueued,
+        avgResponseTimeMs: 0, // Would need tracking
+        errorRate: 0, // Would need tracking
+      });
+    }
+
+    return health;
+  }
+
+  private computeAgentHealth(agents: AgentState[]): AgentHealthEntry[] {
+    return agents.map(agent => ({
+      agentId: agent.id.id,
+      domain: this.getAgentDomain(agent),
+      health: agent.health,
+      status: agent.status,
+      lastHeartbeat: agent.lastHeartbeat,
+      currentLoad: agent.workload,
+      recentErrors: agent.metrics.tasksFailed,
+    }));
+  }
+
+  private detectBottlenecks(
+    status: { domains: DomainStatus[]; metrics: CoordinatorMetrics },
+    agents: AgentState[],
+    metrics: CoordinatorMetrics
+  ): Bottleneck[] {
+    const bottlenecks: Bottleneck[] = [];
+
+    // Check domain queues
+    for (const domain of status.domains) {
+      if (domain.tasksQueued > this.config.bottleneckThresholds.queueDepth) {
+        bottlenecks.push({
+          type: 'domain',
+          location: domain.name,
+          severity: domain.tasksQueued > 20 ? 'critical' : 'high',
+          description: `High task queue depth in ${domain.name} domain`,
+          impact: `${domain.tasksQueued} tasks waiting for processing`,
+          suggestedAction: `Consider scaling ${domain.name} domain agents`,
+        });
+      }
+    }
+
+    // Check error agents
+    const errorAgents = agents.filter(a => a.status === 'error');
+    if (errorAgents.length > 0) {
+      bottlenecks.push({
+        type: 'agent',
+        location: errorAgents.map(a => a.id.id).join(', '),
+        severity: errorAgents.length > 3 ? 'critical' : 'medium',
+        description: `${errorAgents.length} agents in error state`,
+        impact: 'Reduced processing capacity',
+        suggestedAction: 'Investigate and recover error agents',
+      });
+    }
+
+    // Check coordination latency
+    if (metrics.coordinationLatencyMs > this.config.bottleneckThresholds.responseTimeMs) {
+      bottlenecks.push({
+        type: 'resource',
+        location: 'coordination',
+        severity: 'high',
+        description: 'High coordination latency detected',
+        impact: `Current latency: ${metrics.coordinationLatencyMs}ms`,
+        suggestedAction: 'Optimize coordination or reduce concurrent tasks',
+      });
+    }
+
+    return bottlenecks;
+  }
+
+  private generateAlerts(
+    bottlenecks: Bottleneck[],
+    agentHealth: AgentHealthEntry[],
+    metrics: CoordinatorMetrics
+  ): HealthAlert[] {
+    const alerts: HealthAlert[] = [];
+
+    // Generate alerts for critical bottlenecks
+    for (const bottleneck of bottlenecks) {
+      if (bottleneck.severity === 'critical' || bottleneck.severity === 'high') {
+        alerts.push({
+          alertId: `alert_${Date.now()}_${alerts.length}`,
+          type: bottleneck.severity === 'critical' ? 'critical' : 'error',
+          source: bottleneck.location,
+          message: bottleneck.description,
+          timestamp: new Date(),
+          acknowledged: false,
+        });
+      }
+    }
+
+    // Alert on low overall agent health
+    const avgHealth = agentHealth.length > 0
+      ? agentHealth.reduce((sum, a) => sum + a.health, 0) / agentHealth.length
+      : 1.0;
+
+    if (avgHealth < 0.5) {
+      alerts.push({
+        alertId: `alert_${Date.now()}_health`,
+        type: avgHealth < 0.3 ? 'critical' : 'warning',
+        source: 'swarm',
+        message: `Low average agent health: ${(avgHealth * 100).toFixed(1)}%`,
+        timestamp: new Date(),
+        acknowledged: false,
+      });
+    }
+
+    // Alert on high failure rate
+    const failureRate = metrics.totalTasks > 0
+      ? metrics.failedTasks / metrics.totalTasks
+      : 0;
+
+    if (failureRate > this.config.bottleneckThresholds.errorRate) {
+      alerts.push({
+        alertId: `alert_${Date.now()}_failures`,
+        type: failureRate > 0.2 ? 'critical' : 'warning',
+        source: 'tasks',
+        message: `High task failure rate: ${(failureRate * 100).toFixed(1)}%`,
+        timestamp: new Date(),
+        acknowledged: false,
+      });
+    }
+
+    return alerts;
+  }
+
+  private calculateOverallHealth(
+    domainHealth: Map<AgentDomain, DomainHealthStatus>,
+    agentHealth: AgentHealthEntry[],
+    bottlenecks: Bottleneck[]
+  ): number {
+    // Average domain health
+    const domainHealthAvg = domainHealth.size > 0
+      ? Array.from(domainHealth.values()).reduce((sum, d) => sum + d.health, 0) / domainHealth.size
+      : 1.0;
+
+    // Average agent health
+    const agentHealthAvg = agentHealth.length > 0
+      ? agentHealth.reduce((sum, a) => sum + a.health, 0) / agentHealth.length
+      : 1.0;
+
+    // Bottleneck penalty
+    const bottleneckPenalty = bottlenecks.reduce((penalty, b) => {
+      switch (b.severity) {
+        case 'critical': return penalty + 0.2;
+        case 'high': return penalty + 0.1;
+        case 'medium': return penalty + 0.05;
+        case 'low': return penalty + 0.02;
+        default: return penalty;
+      }
+    }, 0);
+
+    const overallHealth = (domainHealthAvg * 0.4 + agentHealthAvg * 0.4) - bottleneckPenalty;
+    return Math.max(0, Math.min(1, overallHealth));
+  }
+
+  private generateRecommendations(bottlenecks: Bottleneck[], overallHealth: number): string[] {
+    const recommendations: string[] = [];
+
+    // Add bottleneck-specific recommendations
+    for (const bottleneck of bottlenecks) {
+      recommendations.push(bottleneck.suggestedAction);
+    }
+
+    // General recommendations based on health
+    if (overallHealth < 0.5) {
+      recommendations.push('Consider reducing task load or adding more agents');
+    }
+
+    if (overallHealth < 0.3) {
+      recommendations.push('URGENT: Investigate system-wide issues immediately');
+    }
+
+    return [...new Set(recommendations)]; // Deduplicate
+  }
+
+  private startHealthMonitoring(): void {
+    this.healthCheckInterval = setInterval(async () => {
+      try {
+        await this.monitorSwarmHealth();
+      } catch (error) {
+        this.emitEvent('queen.health.error', { error: String(error) });
+      }
+    }, this.config.healthCheckIntervalMs);
+  }
+
+  private stopHealthMonitoring(): void {
+    if (this.healthCheckInterval) {
+      clearInterval(this.healthCheckInterval);
+      this.healthCheckInterval = undefined;
+    }
+  }
+
+  // ===========================================================================
+  // Consensus Coordination
+  // ===========================================================================
+
+  /**
+   * Coordinate consensus for a decision
+   *
+   * @param decision - Decision requiring consensus
+   * @returns Consensus result
+   */
+  async coordinateConsensus(decision: Decision): Promise<ConsensusResult> {
+    const startTime = performance.now();
+
+    this.decisionCounter++;
+    decision.decisionId = `decision_${Date.now()}_${this.decisionCounter}`;
+
+    // Store active decision
+    this.activeDecisions.set(decision.decisionId, decision);
+
+    try {
+      let result: ConsensusResult;
+
+      switch (decision.requiredConsensus) {
+        case 'queen-override':
+          // Queen can make unilateral decisions for certain types
+          result = this.queenOverride(decision);
+          break;
+
+        case 'unanimous':
+          result = await this.unanimousConsensus(decision);
+          break;
+
+        case 'supermajority':
+          result = await this.supermajorityConsensus(decision);
+          break;
+
+        case 'weighted':
+          result = await this.weightedConsensus(decision);
+          break;
+
+        case 'majority':
+        default:
+          result = await this.majorityConsensus(decision);
+      }
+
+      // Record latency
+      const latency = performance.now() - startTime;
+      this.consensusLatencies.push(latency);
+      if (this.consensusLatencies.length > 100) {
+        this.consensusLatencies.shift();
+      }
+
+      this.emitEvent('queen.consensus.completed', {
+        decisionId: decision.decisionId,
+        type: decision.requiredConsensus,
+        approved: result.approved,
+        approvalRate: result.approvalRate,
+        latencyMs: latency,
+      });
+
+      return result;
+    } finally {
+      this.activeDecisions.delete(decision.decisionId);
+    }
+  }
+
+  private queenOverride(decision: Decision): ConsensusResult {
+    // Queen can make immediate decisions for:
+    // - Emergency actions
+    // - Agent termination
+    // - Priority overrides
+    const allowedTypes: DecisionType[] = ['emergency-action', 'agent-termination', 'priority-override'];
+
+    if (!allowedTypes.includes(decision.type)) {
+      throw new Error(`Queen override not allowed for decision type: ${decision.type}`);
+    }
+
+    return {
+      proposalId: decision.decisionId,
+      approved: true,
+      approvalRate: 1.0,
+      participationRate: 1.0,
+      finalValue: decision.proposal,
+      rounds: 1,
+      durationMs: 0,
+    };
+  }
+
+  private async majorityConsensus(decision: Decision): Promise<ConsensusResult> {
+    // Use swarm's consensus engine with majority threshold
+    const result = await this.swarm.proposeConsensus({
+      decision,
+      threshold: 0.51,
+      timeout: this.config.consensusTimeouts.majority,
+    });
+
+    return result;
+  }
+
+  private async supermajorityConsensus(decision: Decision): Promise<ConsensusResult> {
+    // Use swarm's consensus engine with 2/3 threshold
+    const result = await this.swarm.proposeConsensus({
+      decision,
+      threshold: 0.67,
+      timeout: this.config.consensusTimeouts.supermajority,
+    });
+
+    return result;
+  }
+
+  private async unanimousConsensus(decision: Decision): Promise<ConsensusResult> {
+    // Use swarm's consensus engine with unanimous requirement
+    const result = await this.swarm.proposeConsensus({
+      decision,
+      threshold: 1.0,
+      timeout: this.config.consensusTimeouts.unanimous,
+    });
+
+    return result;
+  }
+
+  private async weightedConsensus(decision: Decision): Promise<ConsensusResult> {
+    // Weighted consensus based on agent performance
+    // For now, delegate to the standard consensus with metadata
+    const agents = this.swarm.getAllAgents();
+    const weights = new Map<string, number>();
+
+    for (const agent of agents) {
+      const weight = agent.metrics.successRate * agent.health;
+      weights.set(agent.id.id, weight);
+    }
+
+    const result = await this.swarm.proposeConsensus({
+      decision,
+      weights: Object.fromEntries(weights),
+      threshold: 0.51,
+      timeout: this.config.consensusTimeouts.majority,
+    });
+
+    return result;
+  }
+
+  // ===========================================================================
+  // Learning from Outcomes
+  // ===========================================================================
+
+  /**
+   * Record task outcome for learning
+   *
+   * @param task - Completed task
+   * @param result - Task result
+   */
+  async recordOutcome(task: TaskDefinition, result: TaskResult): Promise<void> {
+    // Store in outcome history
+    this.outcomeHistory.push(result);
+    if (this.outcomeHistory.length > 1000) {
+      this.outcomeHistory.shift();
+    }
+
+    // Learn from outcome if neural system available
+    if (this.neural && this.config.enableLearning) {
+      await this.learnFromOutcome(task, result);
+    }
+
+    // Store in memory service if available
+    if (this.memory) {
+      await this.storeOutcomeMemory(task, result);
+    }
+
+    this.emitEvent('queen.outcome.recorded', {
+      taskId: task.id.id,
+      success: result.success,
+      durationMs: result.durationMs,
+      qualityScore: result.metrics.qualityScore,
+    });
+  }
+
+  private async learnFromOutcome(task: TaskDefinition, result: TaskResult): Promise<void> {
+    if (!this.neural) return;
+
+    // Create trajectory for this task
+    const trajectoryId = this.neural.beginTask(
+      task.description || task.name,
+      result.domain
+    );
+
+    // Record the execution step
+    const embedding = this.createSimpleEmbedding(task.description || task.name);
+    const reward = result.success
+      ? result.metrics.qualityScore * 0.8 + 0.2
+      : result.metrics.qualityScore * 0.3;
+
+    this.neural.recordStep(
+      trajectoryId,
+      `executed_${task.type}_in_${result.domain}`,
+      reward,
+      embedding
+    );
+
+    // Complete the task trajectory
+    await this.neural.completeTask(trajectoryId, result.metrics.qualityScore);
+  }
+
+  private async storeOutcomeMemory(task: TaskDefinition, result: TaskResult): Promise<void> {
+    if (!this.memory) return;
+
+    try {
+      await this.memory.store({
+        key: `outcome_${task.id.id}`,
+        content: this.formatOutcomeContent(task, result),
+        namespace: 'queen-outcomes',
+        tags: [
+          task.type,
+          result.domain,
+          result.success ? 'success' : 'failure',
+        ],
+        metadata: {
+          taskId: task.id.id,
+          success: result.success,
+          durationMs: result.durationMs,
+          qualityScore: result.metrics.qualityScore,
+          agentId: result.agentId,
+        },
+      });
+    } catch (error) {
+      // Log but don't fail - memory storage is optional
+      this.emitEvent('queen.memory.error', { error: String(error) });
+    }
+  }
+
+  private formatOutcomeContent(task: TaskDefinition, result: TaskResult): string {
+    return [
+      `Task: ${task.name}`,
+      `Type: ${task.type}`,
+      `Domain: ${result.domain}`,
+      `Agent: ${result.agentId}`,
+      `Success: ${result.success}`,
+      `Duration: ${result.durationMs}ms`,
+      `Quality: ${result.metrics.qualityScore}`,
+      `Description: ${task.description || 'N/A'}`,
+    ].join('\n');
+  }
+
+  // ===========================================================================
+  // Utility Methods
+  // ===========================================================================
+
+  /**
+   * Get the last health report
+   */
+  getLastHealthReport(): HealthReport | undefined {
+    return this.lastHealthReport;
+  }
+
+  /**
+   * Get outcome history
+   */
+  getOutcomeHistory(): TaskResult[] {
+    return [...this.outcomeHistory];
+  }
+
+  /**
+   * Get analysis cache
+   */
+  getAnalysisCache(): Map<string, TaskAnalysis> {
+    return new Map(this.analysisCache);
+  }
+
+  /**
+   * Get delegation plans
+   */
+  getDelegationPlans(): Map<string, DelegationPlan> {
+    return new Map(this.delegationPlans);
+  }
+
+  /**
+   * Get performance statistics
+   */
+  getPerformanceStats(): {
+    avgAnalysisLatencyMs: number;
+    avgDelegationLatencyMs: number;
+    avgConsensusLatencyMs: number;
+    totalAnalyses: number;
+    totalDelegations: number;
+    totalDecisions: number;
+  } {
+    const avg = (arr: number[]) => arr.length > 0
+      ? arr.reduce((a, b) => a + b, 0) / arr.length
+      : 0;
+
+    return {
+      avgAnalysisLatencyMs: avg(this.analysisLatencies),
+      avgDelegationLatencyMs: avg(this.delegationLatencies),
+      avgConsensusLatencyMs: avg(this.consensusLatencies),
+      totalAnalyses: this.analysisCounter,
+      totalDelegations: this.planCounter,
+      totalDecisions: this.decisionCounter,
+    };
+  }
+
+  /**
+   * Check if learning is enabled
+   */
+  isLearningEnabled(): boolean {
+    return this.config.enableLearning && !!this.neural;
+  }
+
+  private emitEvent(type: string, data: Record<string, unknown>): void {
+    const event = {
+      id: `event_${Date.now()}_${Math.random().toString(36).substr(2, 6)}`,
+      type,
+      source: 'queen-coordinator',
+      timestamp: new Date(),
+      data,
+    };
+
+    this.emit(type, event);
+    this.emit('event', event);
+  }
+}
+
+// =============================================================================
+// Factory Function
+// =============================================================================
+
+/**
+ * Create a Queen Coordinator instance
+ */
+export function createQueenCoordinator(
+  swarm: ISwarmCoordinator,
+  config?: Partial<QueenCoordinatorConfig>,
+  neural?: INeuralLearningSystem,
+  memory?: IMemoryService
+): QueenCoordinator {
+  return new QueenCoordinator(swarm, config, neural, memory);
+}
+
+export default QueenCoordinator;