Agent Swarm Testing

Test all three coordination strategies with real LLM execution

This page creates and executes test swarms to verify sequential, parallel, and hierarchical coordination strategies work correctly. Each test creates a new swarm and executes it with a predefined task.

Research Report Generatorsequential

Sequential swarm that researches, analyzes, and writes reports

Agents (3)

Research AgentAnalysis AgentWriting Agent

Test Task

Create a comprehensive report on the current state of multi-agent AI systems and their coordination strategies

Expected Behavior

Each agent builds on previous output sequentially

Multi-Source Code Reviewparallel

Parallel swarm reviewing different aspects simultaneously

Agents (3)

Security ReviewerPerformance ReviewerStyle Reviewer

Test Task

Review this code: ```javascript function processUserData(data) { const users = JSON.parse(data); let result = []; for (let i = 0; i < users.length; i++) { if (users[i].age > 18) { result.push(users[i]); } } return result; } ```

Expected Behavior

All agents execute concurrently, results aggregated

Product Launch Campaignhierarchical

Hierarchical swarm with coordinator and specialists

Agents (4)

Campaign CoordinatorContent CreatorSocial Media StrategistEmail Marketer

Test Task

Create a complete product launch campaign for FlowMaster, an AI-powered productivity tool

Expected Behavior

Coordinator delegates to workers and synthesizes results