Startup Agent

Summary

The Startup Agent is a lifecycle management agent that controls agent startup and initialization for the mindX system. It manages the startup sequence, initializes always-on agents, loads agent registry from pgvectorscale, and restores agent state from blockchain if needed.

Technical Explanation

The Startup Agent orchestrates the system startup process, ensuring that all components are initialized in the correct order and that agent state is properly restored from persistent storage.

Architecture

• Type: lifecycle_management_agent
• Location: agents/orchestration/startup_agent.py
• Lifecycle: Always-on (critical system component)
• Integration: Coordinates with CoordinatorAgent, MemoryAgent, IDManagerAgent

Core Capabilities

• System Initialization: Initializes the entire mindX system
• Agent Registry Loading: Loads agent registry from pgvectorscale
• Blockchain State Restoration: Restores agent state from blockchain if needed
• Always-On Agent Initialization: Initializes always-on agents (Coordinator, mindXagent, MemoryAgent)
• On-Demand Agent Initialization: Initializes agents on-demand when requested
• Startup Sequence Coordination: Coordinates the startup sequence

Always-On Agents

The following agents are always initialized:

• coordinator_agent: Central orchestrator
• mindxagent: System overseer
• memory_agent: Memory management

Startup Sequence

1. Load agent registry from pgvectorscale
2. Restore agent state from blockchain (if requested)
3. Initialize always-on agents
4. Coordinate startup sequence
5. System ready

Usage

from agents.orchestration.startup_agent import StartupAgent
from agents.orchestration.coordinator_agent import CoordinatorAgent
from agents.memory_agent import MemoryAgent

Create startup agent
startup_agent = StartupAgent(
    agent_id="startup_agent",
    coordinator_agent=coordinator_agent,
    memory_agent=memory_agent
)

Initialize system
initialization_result = await startup_agent.initialize_system(
    restore_from_blockchain=False
)

Initialize agent on-demand
on_demand_result = await startup_agent.initialize_agent_on_demand(
    agent_type="prediction_agent",
    agent_id="prediction_001",
    config={"time_horizon": "24h"}
)

Integration with MastermindAgent

The Startup Agent is initialized by MastermindAgent during system startup:

# MastermindAgent initializes lifecycle agents
await mastermind_agent._initialize_lifecycle_agents()
Startup agent is available in mastermind_agent.lifecycle_agents["startup"]

Integration with ReplicationAgent

The Startup Agent works with ReplicationAgent to restore agent state:

• Loads agent registry from pgvectorscale
• Restores agent state from blockchain if needed
• Coordinates with ReplicationAgent for state restoration

NFT Metadata (iNFT/dNFT Ready)

iNFT (Intelligent NFT) Metadata

{
  "name": "mindX Startup Agent",
  "description": "Lifecycle management agent for system startup and initialization",
  "image": "ipfs://[avatar_cid]",
  "external_url": "https://mindx.internal/agents/orchestration/startup_agent",
  "attributes": [
    {
      "trait_type": "Agent Type",
      "value": "lifecycle_management_agent"
    },
    {
      "trait_type": "Capability",
      "value": "System Initialization"
    },
    {
      "trait_type": "Lifecycle",
      "value": "always_on"
    },
    {
      "trait_type": "Critical Component",
      "value": "true"
    }
  ]
}

Design Decisions

• Always-On: Critical system component, always initialized
• Sequential Initialization: Ensures proper initialization order
• State Restoration: Supports blockchain state restoration
• On-Demand Support: Can initialize agents on-demand
• Registry Integration: Loads from pgvectorscale

Future Enhancements

• Parallel initialization where safe
• Health checks during startup
• Startup time optimization
• Startup failure recovery
• Startup metrics and monitoring