CrewAI Agent

Required Parameters

• Agents: Team member configurations
• Tasks: Task definitions and assignments
• Language Model: The AI model for processing
• Process Flow: Task execution order
• Memory System: Shared knowledge base

Optional Configuration

• Tools: Available tools and capabilities
  • name: Tool identifier
  • description: Tool purpose
  • parameters: Required inputs
• Delegation Rules: Task delegation settings
  • allowDelegation: Enable delegation
  • maxDelegationDepth: Maximum chain length
  • delegationStrategy: Assignment strategy
• Monitoring: Performance tracking options

Output Format

{
  "result": {
    "taskResults": [
      {
        "taskId": string,
        "agentId": string,
        "output": any,
        "status": string
      }
    ],
    "metadata": {
      "completionTime": string,
      "agentsInvolved": array,
      "taskSequence": array
    }
  }
}

Example Usage

const crewAIAgent = new CrewAIAgent({
  agents: [
    {
      id: "researcher",
      role: "Research information",
      expertise: ["web-search", "data-analysis"]
    },
    {
      id: "writer",
      role: "Content creation",
      expertise: ["writing", "editing"]
    },
    {
      id: "reviewer",
      role: "Quality assurance",
      expertise: ["fact-checking", "proofreading"]
    }
  ],
  tasks: [
    {
      id: "research",
      assignedTo: "researcher",
      description: "Gather information"
    },
    {
      id: "write",
      assignedTo: "writer",
      description: "Create content"
    },
    {
      id: "review",
      assignedTo: "reviewer",
      description: "Review and approve"
    }
  ],
  languageModel: "gpt-4",
  processFlow: "sequential",
  memorySystem: "shared"
});

const result = await crewAIAgent.process({
  input: "Create a research article about AI trends"
});