Creating Your Subagent

Step 1: Provide Detailed Prompt

In the chat interface, provide a comprehensive prompt describing what your subagent should do. Be specific about:

Fig 9.2

Step 2: Agent Creation Process Begins

Once you submit the prompt, the Agent Hive will:

• Parse your requirements
• Analyze the capabilities needed
• Begin generating the agent configuration
• Display progress in real-time

Step 3: User Approval

The system includes human-in-the-loop capabilities for quality control.

What You’ll See:

• Questions about specific requirements

• Tool permission requests

• Configuration choices

  
  

  

  Fig 9.3

Your Action: Review and approve each step by clicking “Submit” or providing your choice

Step 4: Agent Configuration Generated

After your input, the agent creates a complete configuration including:

Configuration Details Displayed:

• Capabilities: What the agent can do (e.g., “Data analysis using Python, R, and SQL”)
• Boundaries: What the agent should avoid
• Methodology: Problem-solving approach
• Tools: Configured tools with permissions (e.g., code_execution_tool, sql tool)
• Tool Permissions: Mode settings (approval_required, disallowed, etc.)

Step 5: Agent Created Successfully✅

Once creation is complete, you’ll see:

Confirmation Message:

Fig 9.4

What This Means:

Agent is now available in your system

Step 5: Verify in Subagents Section

Navigate to the Subagents section to verify:

• Click on the agent to view full configuration

• Click the “Run” button to activate the agent (turns it active/ready to use)

  
  

  

  Fig 9.5