agentic-flow 1.9.0 → 1.9.1

package/bench/benchmark.js

@@ -1,244 +0,0 @@
-#!/usr/bin/env node
-/**
- * ReasoningBank Benchmark Orchestrator
- *
- * Runs comprehensive benchmarks comparing baseline vs ReasoningBank agents
- */
-import { BaselineAgent } from './agents/baseline-agent.js';
-import { ReasoningBankAgent } from './agents/reasoningbank-agent.js';
-import { MetricsCollector } from './lib/metrics.js';
-import { generateReport } from './lib/report-generator.js';
-import { codingTasksScenario } from './scenarios/coding-tasks.js';
-import { debuggingTasksScenario } from './scenarios/debugging-tasks.js';
-import { apiDesignTasksScenario } from './scenarios/api-design-tasks.js';
-import { problemSolvingTasksScenario } from './scenarios/problem-solving-tasks.js';
-import { db } from '../agentic-flow/src/reasoningbank/index.js';
-// Default configuration
-const DEFAULT_CONFIG = {
-    iterations: 3,
-    parallelAgents: 1,
-    scenarios: ['coding-tasks', 'debugging-tasks', 'api-design-tasks', 'problem-solving-tasks'],
-    enableWarmStart: false,
-    memorySize: 1000,
-    outputFormats: ['json', 'markdown'],
-    anthropicApiKey: process.env.ANTHROPIC_API_KEY
-};
-class BenchmarkOrchestrator {
-    config;
-    scenarios;
-    results = [];
-    constructor(config = {}) {
-        this.config = { ...DEFAULT_CONFIG, ...config };
-        this.scenarios = new Map([
-            ['coding-tasks', codingTasksScenario],
-            ['debugging-tasks', debuggingTasksScenario],
-            ['api-design-tasks', apiDesignTasksScenario],
-            ['problem-solving-tasks', problemSolvingTasksScenario]
-        ]);
-    }
-    async run() {
-        console.log('\n🧪 ReasoningBank Benchmark Suite');
-        console.log('═'.repeat(60));
-        console.log(`Scenarios: ${this.config.scenarios.join(', ')}`);
-        console.log(`Iterations: ${this.config.iterations}`);
-        console.log(`Warm Start: ${this.config.enableWarmStart ? 'Yes' : 'No'}\n`);
-        const startTime = Date.now();
-        // Initialize database
-        console.log('📦 Initializing database...');
-        await db.runMigrations();
-        // Run each scenario
-        for (const scenarioName of this.config.scenarios) {
-            const scenario = this.scenarios.get(scenarioName);
-            if (!scenario) {
-                console.error(`❌ Unknown scenario: ${scenarioName}`);
-                continue;
-            }
-            const result = await this.runScenario(scenario);
-            this.results.push(result);
-        }
-        const executionTime = Date.now() - startTime;
-        // Generate report
-        const report = this.generateBenchmarkReport(executionTime);
-        // Save results
-        await this.saveResults(report);
-        // Print summary
-        this.printSummary(report);
-        return report;
-    }
-    async runScenario(scenario) {
-        console.log(`\n🎯 Running Scenario: ${scenario.name}`);
-        console.log('─'.repeat(60));
-        const metrics = new MetricsCollector();
-        // Create agents
-        const baseline = new BaselineAgent(this.config.anthropicApiKey);
-        const reasoningbank = new ReasoningBankAgent(this.config.anthropicApiKey);
-        // Reset agents
-        await baseline.reset();
-        await reasoningbank.reset();
-        // Run iterations
-        for (let iteration = 0; iteration < this.config.iterations; iteration++) {
-            console.log(`\n  Iteration ${iteration + 1}/${this.config.iterations}`);
-            let baselineSuccesses = 0;
-            let reasoningbankSuccesses = 0;
-            let baselineTokens = 0;
-            let reasoningbankTokens = 0;
-            // Run all tasks in this iteration
-            for (let i = 0; i < scenario.tasks.length; i++) {
-                const task = scenario.tasks[i];
-                const progress = ((i + 1) / scenario.tasks.length * 100).toFixed(0);
-                process.stdout.write(`\r    Task ${i + 1}/${scenario.tasks.length} (${progress}%)`);
-                // Run baseline
-                const baselineResult = await baseline.executeTask(task);
-                metrics.addResult(baselineResult);
-                if (baselineResult.success)
-                    baselineSuccesses++;
-                baselineTokens += baselineResult.tokens;
-                // Run ReasoningBank
-                const rbResult = await reasoningbank.executeTask(task);
-                metrics.addResult(rbResult);
-                if (rbResult.success)
-                    reasoningbankSuccesses++;
-                reasoningbankTokens += rbResult.tokens;
-            }
-            // Record learning point
-            const totalTasks = scenario.tasks.length;
-            const memoryStats = await reasoningbank.getMemoryStats();
-            metrics.addLearningPoint({
-                iteration: iteration + 1,
-                baselineSuccess: baselineSuccesses / totalTasks,
-                reasoningbankSuccess: reasoningbankSuccesses / totalTasks,
-                baselineTokens: baselineTokens / totalTasks,
-                reasoningbankTokens: reasoningbankTokens / totalTasks,
-                memoriesAvailable: memoryStats.total
-            });
-            console.log(`\n    ├─ Baseline:      ${baselineSuccesses}/${totalTasks} success (${baselineTokens.toLocaleString()} tokens)`);
-            console.log(`    └─ ReasoningBank: ${reasoningbankSuccesses}/${totalTasks} success (${reasoningbankTokens.toLocaleString()} tokens, ${memoryStats.total} memories)`);
-        }
-        const results = metrics.generateScenarioResults(scenario.name);
-        // Print scenario summary
-        console.log('\n  📊 Scenario Results:');
-        console.log(`    Success Rate: ${(results.baseline.successRate * 100).toFixed(1)}% → ${(results.reasoningbank.successRate * 100).toFixed(1)}% (${results.improvement.successRateDelta})`);
-        console.log(`    Token Efficiency: ${results.improvement.tokenEfficiency}`);
-        console.log(`    Latency Overhead: ${results.improvement.latencyOverhead}`);
-        if (results.improvement.learningVelocity) {
-            console.log(`    Learning Velocity: ${results.improvement.learningVelocity.toFixed(1)}x faster`);
-        }
-        const insights = metrics.generateInsights(results);
-        if (insights.length > 0) {
-            console.log('\n  💡 Insights:');
-            insights.forEach(insight => console.log(`    ${insight}`));
-        }
-        return results;
-    }
-    generateBenchmarkReport(executionTime) {
-        // Calculate overall metrics
-        const totalTasks = this.results.reduce((sum, r) => sum + r.baseline.totalTasks + r.reasoningbank.totalTasks, 0);
-        const overallBaseline = {
-            successRate: 0,
-            avgTokens: 0,
-            avgLatency: 0
-        };
-        const overallRB = {
-            successRate: 0,
-            avgTokens: 0,
-            avgLatency: 0
-        };
-        let totalBaselineTasks = 0;
-        let totalRBTasks = 0;
-        for (const result of this.results) {
-            overallBaseline.successRate += result.baseline.successRate * result.baseline.totalTasks;
-            overallBaseline.avgTokens += result.baseline.avgTokens * result.baseline.totalTasks;
-            overallBaseline.avgLatency += result.baseline.avgLatency * result.baseline.totalTasks;
-            totalBaselineTasks += result.baseline.totalTasks;
-            overallRB.successRate += result.reasoningbank.successRate * result.reasoningbank.totalTasks;
-            overallRB.avgTokens += result.reasoningbank.avgTokens * result.reasoningbank.totalTasks;
-            overallRB.avgLatency += result.reasoningbank.avgLatency * result.reasoningbank.totalTasks;
-            totalRBTasks += result.reasoningbank.totalTasks;
-        }
-        overallBaseline.successRate /= totalBaselineTasks;
-        overallBaseline.avgTokens /= totalBaselineTasks;
-        overallBaseline.avgLatency /= totalBaselineTasks;
-        overallRB.successRate /= totalRBTasks;
-        overallRB.avgTokens /= totalRBTasks;
-        overallRB.avgLatency /= totalRBTasks;
-        const overallImprovement = {
-            successRateDelta: `+${((overallRB.successRate - overallBaseline.successRate) * 100).toFixed(1)}%`,
-            tokenEfficiency: `+${((overallBaseline.avgTokens - overallRB.avgTokens) / overallBaseline.avgTokens * 100).toFixed(1)}%`,
-            latencyOverhead: `+${((overallRB.avgLatency - overallBaseline.avgLatency) / overallBaseline.avgLatency * 100).toFixed(1)}%`
-        };
-        // Generate recommendations
-        const recommendations = this.generateRecommendations();
-        return {
-            summary: {
-                totalScenarios: this.results.length,
-                totalTasks,
-                overallImprovement,
-                executionTime
-            },
-            scenarios: this.results,
-            recommendations,
-            generatedAt: new Date().toISOString()
-        };
-    }
-    generateRecommendations() {
-        const recommendations = [];
-        for (const result of this.results) {
-            if (result.improvement.successRatePercent < 0) {
-                recommendations.push(`⚠️  ${result.scenarioName}: Baseline outperformed ReasoningBank - review memory quality`);
-            }
-            if (result.improvement.tokenSavings < 0) {
-                recommendations.push(`💡 ${result.scenarioName}: Token overhead detected - optimize memory retrieval (k parameter)`);
-            }
-            if (result.improvement.latencyDelta > 1000) {
-                recommendations.push(`⚡ ${result.scenarioName}: High latency - consider caching or reducing consolidation frequency`);
-            }
-        }
-        if (recommendations.length === 0) {
-            recommendations.push('✅ All metrics look good! ReasoningBank is performing optimally.');
-        }
-        return recommendations;
-    }
-    async saveResults(report) {
-        const fs = await import('fs/promises');
-        const path = await import('path');
-        // Save JSON
-        if (this.config.outputFormats.includes('json')) {
-            const jsonPath = path.join(process.cwd(), 'bench', 'results', `benchmark-${Date.now()}.json`);
-            await fs.writeFile(jsonPath, JSON.stringify(report, null, 2));
-            console.log(`\n📄 JSON report saved: ${jsonPath}`);
-        }
-        // Save Markdown
-        if (this.config.outputFormats.includes('markdown')) {
-            const mdPath = path.join(process.cwd(), 'bench', 'reports', `benchmark-${Date.now()}.md`);
-            const markdown = generateReport(report);
-            await fs.writeFile(mdPath, markdown);
-            console.log(`📄 Markdown report saved: ${mdPath}`);
-        }
-    }
-    printSummary(report) {
-        console.log('\n\n' + '═'.repeat(60));
-        console.log('📊 BENCHMARK SUMMARY');
-        console.log('═'.repeat(60));
-        console.log(`Total Scenarios: ${report.summary.totalScenarios}`);
-        console.log(`Total Tasks: ${report.summary.totalTasks}`);
-        console.log(`Execution Time: ${(report.summary.executionTime / 1000).toFixed(1)}s`);
-        console.log('\nOverall Improvement:');
-        console.log(`  Success Rate: ${report.summary.overallImprovement.successRateDelta}`);
-        console.log(`  Token Efficiency: ${report.summary.overallImprovement.tokenEfficiency}`);
-        console.log(`  Latency Overhead: ${report.summary.overallImprovement.latencyOverhead}`);
-        if (report.recommendations.length > 0) {
-            console.log('\n💡 Recommendations:');
-            report.recommendations.forEach(rec => console.log(`  ${rec}`));
-        }
-        console.log('\n' + '═'.repeat(60) + '\n');
-    }
-}
-// CLI Entry Point
-if (import.meta.url === `file://${process.argv[1]}`) {
-    const orchestrator = new BenchmarkOrchestrator();
-    orchestrator.run().catch(error => {
-        console.error('\n❌ Benchmark failed:', error);
-        process.exit(1);
-    });
-}
-export { BenchmarkOrchestrator };

package/bench/benchmark.ts

@@ -1,309 +0,0 @@
-#!/usr/bin/env node
-/**
- * ReasoningBank Benchmark Orchestrator
- *
- * Runs comprehensive benchmarks comparing baseline vs ReasoningBank agents
- */
-
-import { BaselineAgent } from './agents/baseline-agent.js';
-import { ReasoningBankAgent } from './agents/reasoningbank-agent.js';
-import { MetricsCollector } from './lib/metrics.js';
-import { generateReport } from './lib/report-generator.js';
-import { codingTasksScenario } from './scenarios/coding-tasks.js';
-import { debuggingTasksScenario } from './scenarios/debugging-tasks.js';
-import { apiDesignTasksScenario } from './scenarios/api-design-tasks.js';
-import { problemSolvingTasksScenario } from './scenarios/problem-solving-tasks.js';
-import { db } from '../agentic-flow/src/reasoningbank/index.js';
-import type { BenchmarkScenario, BenchmarkConfig, ScenarioResults, BenchmarkReport } from './lib/types.js';
-
-// Default configuration
-const DEFAULT_CONFIG: BenchmarkConfig = {
-  iterations: 3,
-  parallelAgents: 1,
-  scenarios: ['coding-tasks', 'debugging-tasks', 'api-design-tasks', 'problem-solving-tasks'],
-  enableWarmStart: false,
-  memorySize: 1000,
-  outputFormats: ['json', 'markdown'],
-  anthropicApiKey: process.env.ANTHROPIC_API_KEY
-};
-
-class BenchmarkOrchestrator {
-  private config: BenchmarkConfig;
-  private scenarios: Map<string, BenchmarkScenario>;
-  private results: ScenarioResults[] = [];
-
-  constructor(config: Partial<BenchmarkConfig> = {}) {
-    this.config = { ...DEFAULT_CONFIG, ...config };
-    this.scenarios = new Map([
-      ['coding-tasks', codingTasksScenario],
-      ['debugging-tasks', debuggingTasksScenario],
-      ['api-design-tasks', apiDesignTasksScenario],
-      ['problem-solving-tasks', problemSolvingTasksScenario]
-    ]);
-  }
-
-  async run(): Promise<BenchmarkReport> {
-    console.log('\n🧪 ReasoningBank Benchmark Suite');
-    console.log('═'.repeat(60));
-    console.log(`Scenarios: ${this.config.scenarios.join(', ')}`);
-    console.log(`Iterations: ${this.config.iterations}`);
-    console.log(`Warm Start: ${this.config.enableWarmStart ? 'Yes' : 'No'}\n`);
-
-    const startTime = Date.now();
-
-    // Initialize database
-    console.log('📦 Initializing database...');
-    await db.runMigrations();
-
-    // Run each scenario
-    for (const scenarioName of this.config.scenarios) {
-      const scenario = this.scenarios.get(scenarioName);
-      if (!scenario) {
-        console.error(`❌ Unknown scenario: ${scenarioName}`);
-        continue;
-      }
-
-      const result = await this.runScenario(scenario);
-      this.results.push(result);
-    }
-
-    const executionTime = Date.now() - startTime;
-
-    // Generate report
-    const report = this.generateBenchmarkReport(executionTime);
-
-    // Save results
-    await this.saveResults(report);
-
-    // Print summary
-    this.printSummary(report);
-
-    return report;
-  }
-
-  private async runScenario(scenario: BenchmarkScenario): Promise<ScenarioResults> {
-    console.log(`\n🎯 Running Scenario: ${scenario.name}`);
-    console.log('─'.repeat(60));
-
-    const metrics = new MetricsCollector();
-
-    // Create agents
-    const baseline = new BaselineAgent(this.config.anthropicApiKey);
-    const reasoningbank = new ReasoningBankAgent(this.config.anthropicApiKey);
-
-    // Reset agents
-    await baseline.reset();
-    await reasoningbank.reset();
-
-    // Run iterations
-    for (let iteration = 0; iteration < this.config.iterations; iteration++) {
-      console.log(`\n  Iteration ${iteration + 1}/${this.config.iterations}`);
-
-      let baselineSuccesses = 0;
-      let reasoningbankSuccesses = 0;
-      let baselineTokens = 0;
-      let reasoningbankTokens = 0;
-
-      // Run all tasks in this iteration
-      for (let i = 0; i < scenario.tasks.length; i++) {
-        const task = scenario.tasks[i];
-        const progress = ((i + 1) / scenario.tasks.length * 100).toFixed(0);
-
-        process.stdout.write(`\r    Task ${i + 1}/${scenario.tasks.length} (${progress}%)`);
-
-        // Run baseline
-        const baselineResult = await baseline.executeTask(task);
-        metrics.addResult(baselineResult);
-        if (baselineResult.success) baselineSuccesses++;
-        baselineTokens += baselineResult.tokens;
-
-        // Run ReasoningBank
-        const rbResult = await reasoningbank.executeTask(task);
-        metrics.addResult(rbResult);
-        if (rbResult.success) reasoningbankSuccesses++;
-        reasoningbankTokens += rbResult.tokens;
-      }
-
-      // Record learning point
-      const totalTasks = scenario.tasks.length;
-      const memoryStats = await reasoningbank.getMemoryStats!();
-
-      metrics.addLearningPoint({
-        iteration: iteration + 1,
-        baselineSuccess: baselineSuccesses / totalTasks,
-        reasoningbankSuccess: reasoningbankSuccesses / totalTasks,
-        baselineTokens: baselineTokens / totalTasks,
-        reasoningbankTokens: reasoningbankTokens / totalTasks,
-        memoriesAvailable: memoryStats.total
-      });
-
-      console.log(`\n    ├─ Baseline:      ${baselineSuccesses}/${totalTasks} success (${baselineTokens.toLocaleString()} tokens)`);
-      console.log(`    └─ ReasoningBank: ${reasoningbankSuccesses}/${totalTasks} success (${reasoningbankTokens.toLocaleString()} tokens, ${memoryStats.total} memories)`);
-    }
-
-    const results = metrics.generateScenarioResults(scenario.name);
-
-    // Print scenario summary
-    console.log('\n  📊 Scenario Results:');
-    console.log(`    Success Rate: ${(results.baseline.successRate * 100).toFixed(1)}% → ${(results.reasoningbank.successRate * 100).toFixed(1)}% (${results.improvement.successRateDelta})`);
-    console.log(`    Token Efficiency: ${results.improvement.tokenEfficiency}`);
-    console.log(`    Latency Overhead: ${results.improvement.latencyOverhead}`);
-
-    if (results.improvement.learningVelocity) {
-      console.log(`    Learning Velocity: ${results.improvement.learningVelocity.toFixed(1)}x faster`);
-    }
-
-    const insights = metrics.generateInsights(results);
-    if (insights.length > 0) {
-      console.log('\n  💡 Insights:');
-      insights.forEach(insight => console.log(`    ${insight}`));
-    }
-
-    return results;
-  }
-
-  private generateBenchmarkReport(executionTime: number): BenchmarkReport {
-    // Calculate overall metrics
-    const totalTasks = this.results.reduce(
-      (sum, r) => sum + r.baseline.totalTasks + r.reasoningbank.totalTasks,
-      0
-    );
-
-    const overallBaseline = {
-      successRate: 0,
-      avgTokens: 0,
-      avgLatency: 0
-    };
-
-    const overallRB = {
-      successRate: 0,
-      avgTokens: 0,
-      avgLatency: 0
-    };
-
-    let totalBaselineTasks = 0;
-    let totalRBTasks = 0;
-
-    for (const result of this.results) {
-      overallBaseline.successRate += result.baseline.successRate * result.baseline.totalTasks;
-      overallBaseline.avgTokens += result.baseline.avgTokens * result.baseline.totalTasks;
-      overallBaseline.avgLatency += result.baseline.avgLatency * result.baseline.totalTasks;
-      totalBaselineTasks += result.baseline.totalTasks;
-
-      overallRB.successRate += result.reasoningbank.successRate * result.reasoningbank.totalTasks;
-      overallRB.avgTokens += result.reasoningbank.avgTokens * result.reasoningbank.totalTasks;
-      overallRB.avgLatency += result.reasoningbank.avgLatency * result.reasoningbank.totalTasks;
-      totalRBTasks += result.reasoningbank.totalTasks;
-    }
-
-    overallBaseline.successRate /= totalBaselineTasks;
-    overallBaseline.avgTokens /= totalBaselineTasks;
-    overallBaseline.avgLatency /= totalBaselineTasks;
-
-    overallRB.successRate /= totalRBTasks;
-    overallRB.avgTokens /= totalRBTasks;
-    overallRB.avgLatency /= totalRBTasks;
-
-    const overallImprovement: any = {
-      successRateDelta: `+${((overallRB.successRate - overallBaseline.successRate) * 100).toFixed(1)}%`,
-      tokenEfficiency: `+${((overallBaseline.avgTokens - overallRB.avgTokens) / overallBaseline.avgTokens * 100).toFixed(1)}%`,
-      latencyOverhead: `+${((overallRB.avgLatency - overallBaseline.avgLatency) / overallBaseline.avgLatency * 100).toFixed(1)}%`
-    };
-
-    // Generate recommendations
-    const recommendations = this.generateRecommendations();
-
-    return {
-      summary: {
-        totalScenarios: this.results.length,
-        totalTasks,
-        overallImprovement,
-        executionTime
-      },
-      scenarios: this.results,
-      recommendations,
-      generatedAt: new Date().toISOString()
-    };
-  }
-
-  private generateRecommendations(): string[] {
-    const recommendations: string[] = [];
-
-    for (const result of this.results) {
-      if (result.improvement.successRatePercent < 0) {
-        recommendations.push(
-          `⚠️  ${result.scenarioName}: Baseline outperformed ReasoningBank - review memory quality`
-        );
-      }
-
-      if (result.improvement.tokenSavings < 0) {
-        recommendations.push(
-          `💡 ${result.scenarioName}: Token overhead detected - optimize memory retrieval (k parameter)`
-        );
-      }
-
-      if (result.improvement.latencyDelta > 1000) {
-        recommendations.push(
-          `⚡ ${result.scenarioName}: High latency - consider caching or reducing consolidation frequency`
-        );
-      }
-    }
-
-    if (recommendations.length === 0) {
-      recommendations.push('✅ All metrics look good! ReasoningBank is performing optimally.');
-    }
-
-    return recommendations;
-  }
-
-  private async saveResults(report: BenchmarkReport): Promise<void> {
-    const fs = await import('fs/promises');
-    const path = await import('path');
-
-    // Save JSON
-    if (this.config.outputFormats.includes('json')) {
-      const jsonPath = path.join(process.cwd(), 'bench', 'results', `benchmark-${Date.now()}.json`);
-      await fs.writeFile(jsonPath, JSON.stringify(report, null, 2));
-      console.log(`\n📄 JSON report saved: ${jsonPath}`);
-    }
-
-    // Save Markdown
-    if (this.config.outputFormats.includes('markdown')) {
-      const mdPath = path.join(process.cwd(), 'bench', 'reports', `benchmark-${Date.now()}.md`);
-      const markdown = generateReport(report);
-      await fs.writeFile(mdPath, markdown);
-      console.log(`📄 Markdown report saved: ${mdPath}`);
-    }
-  }
-
-  private printSummary(report: BenchmarkReport): void {
-    console.log('\n\n' + '═'.repeat(60));
-    console.log('📊 BENCHMARK SUMMARY');
-    console.log('═'.repeat(60));
-    console.log(`Total Scenarios: ${report.summary.totalScenarios}`);
-    console.log(`Total Tasks: ${report.summary.totalTasks}`);
-    console.log(`Execution Time: ${(report.summary.executionTime / 1000).toFixed(1)}s`);
-    console.log('\nOverall Improvement:');
-    console.log(`  Success Rate: ${report.summary.overallImprovement.successRateDelta}`);
-    console.log(`  Token Efficiency: ${report.summary.overallImprovement.tokenEfficiency}`);
-    console.log(`  Latency Overhead: ${report.summary.overallImprovement.latencyOverhead}`);
-
-    if (report.recommendations.length > 0) {
-      console.log('\n💡 Recommendations:');
-      report.recommendations.forEach(rec => console.log(`  ${rec}`));
-    }
-
-    console.log('\n' + '═'.repeat(60) + '\n');
-  }
-}
-
-// CLI Entry Point
-if (import.meta.url === `file://${process.argv[1]}`) {
-  const orchestrator = new BenchmarkOrchestrator();
-  orchestrator.run().catch(error => {
-    console.error('\n❌ Benchmark failed:', error);
-    process.exit(1);
-  });
-}
-
-export { BenchmarkOrchestrator };

package/bench/config.json

@@ -1,79 +0,0 @@
-{
-  "benchmark": {
-    "name": "ReasoningBank Comprehensive Benchmark",
-    "version": "1.0.0",
-    "description": "Comprehensive benchmark comparing baseline vs ReasoningBank agents"
-  },
-  "execution": {
-    "iterations": 3,
-    "parallelAgents": 1,
-    "enableWarmStart": false,
-    "memorySize": 1000,
-    "randomSeed": null
-  },
-  "scenarios": [
-    {
-      "name": "coding-tasks",
-      "enabled": true,
-      "iterations": 3,
-      "description": "Tests implementation of common programming patterns"
-    },
-    {
-      "name": "debugging-tasks",
-      "enabled": true,
-      "iterations": 3,
-      "description": "Tests ability to identify and fix bugs"
-    },
-    {
-      "name": "api-design-tasks",
-      "enabled": true,
-      "iterations": 3,
-      "description": "Tests API design and best practices"
-    },
-    {
-      "name": "problem-solving-tasks",
-      "enabled": true,
-      "iterations": 3,
-      "description": "Tests algorithmic problem solving"
-    }
-  ],
-  "agents": {
-    "baseline": {
-      "model": "claude-sonnet-4-5-20250929",
-      "maxTokens": 4096,
-      "temperature": 0.7
-    },
-    "reasoningbank": {
-      "model": "claude-sonnet-4-5-20250929",
-      "maxTokens": 4096,
-      "temperature": 0.7,
-      "memoryConfig": {
-        "k": 3,
-        "alpha": 0.65,
-        "beta": 0.15,
-        "gamma": 0.20,
-        "delta": 0.10,
-        "consolidationThreshold": 100
-      }
-    }
-  },
-  "output": {
-    "formats": ["json", "markdown", "csv"],
-    "directories": {
-      "results": "bench/results",
-      "reports": "bench/reports"
-    },
-    "saveTrajectories": false,
-    "verboseLogging": false
-  },
-  "validation": {
-    "confidenceInterval": 0.95,
-    "minIterations": 3,
-    "significanceThreshold": 0.05
-  },
-  "performance": {
-    "maxConcurrentTasks": 5,
-    "taskTimeout": 60000,
-    "retryFailedTasks": false
-  }
-}