> ## Documentation Index
> Fetch the complete documentation index at: https://docs.argentos.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Agents Overview

> Understand how ArgentOS agents work — the AI runtime at the core of the system.

## What is an Agent?

An ArgentOS agent is a persistent AI runtime that can receive messages, reason about them, use tools to take actions, and respond. Unlike a simple chatbot, an agent has:

* **Persistent memory** across sessions (via MemU)
* **Tool access** to execute commands, browse the web, manage tasks, and more
* **Multi-channel presence** across Telegram, Discord, WhatsApp, and other platforms
* **Autonomous capabilities** via the heartbeat and task system

## Core Architecture

<CardGroup cols={2}>
  <Card title="Agent Types & Roles" icon="users-gear" href="/agents/agent-types">
    Four archetypes: Main, Family, Sub-Agent, and Worker.
  </Card>

  <Card title="Architecture" icon="sitemap" href="/agents/architecture">
    How the runtime connects the gateway, Claude API, tools, and memory.
  </Card>

  <Card title="The Agent Loop" icon="arrows-spin" href="/agents/agent-loop">
    Receive, think, act, respond — the core execution cycle.
  </Card>

  <Card title="Consciousness Kernel" icon="brain" href="/agents/consciousness-kernel">
    The persistent autonomous executive for continuous agency.
  </Card>
</CardGroup>

## Autonomy & Accountability

<CardGroup cols={2}>
  <Card title="Contemplation" icon="thought-bubble" href="/agents/contemplation">
    Self-directed thinking cycles for reflection and growth.
  </Card>

  <Card title="Accountability System" icon="shield-check" href="/agents/accountability">
    Scoring, penalties, rewards, and trust metrics.
  </Card>

  <Card title="Heartbeat Contracts" icon="file-contract" href="/agents/heartbeat-contracts">
    Structured task contracts with verification.
  </Card>

  <Card title="RALF + ANGEL" icon="scale-balanced" href="/agents/ralf-angel">
    Response verification and ground truth checking.
  </Card>
</CardGroup>

## Configuration & Identity

<CardGroup cols={2}>
  <Card title="System Prompt" icon="terminal" href="/agents/system-prompt">
    How the system prompt is assembled from modular sections.
  </Card>

  <Card title="Alignment Documents" icon="scroll" href="/agents/alignment-docs">
    SOUL.md, IDENTITY.md, and other personality-defining docs.
  </Card>

  <Card title="Sessions" icon="comments" href="/agents/sessions">
    Conversation management, context windows, and compaction.
  </Card>

  <Card title="Multi-Agent" icon="people-group" href="/agents/multi-agent">
    Running multiple agents with different configurations.
  </Card>
</CardGroup>

## Advanced Systems

<CardGroup cols={2}>
  <Card title="Intent Governance" icon="gavel" href="/agents/intent-governance">
    Three-tier policy engine constraining agent behavior.
  </Card>

  <Card title="Execution Worker" icon="gears" href="/agents/execution-worker">
    Autonomous background task processor.
  </Card>

  <Card title="Knowledge Library" icon="book" href="/agents/knowledge-library">
    PostgreSQL-backed RAG with agent-level ACL.
  </Card>

  <Card title="Self-Improving System" icon="chart-line" href="/agents/sis-architecture">
    How ArgentOS learns from its own behavior.
  </Card>

  <Card title="SpecForge" icon="wand-magic-sparkles" href="/agents/specforge">
    5-stage project management from intent to execution.
  </Card>

  <Card title="Argent Runtime" icon="microchip" href="/agents/argent-runtime">
    The native agent runtime and LLM layer.
  </Card>

  <Card title="Outbound Messaging" icon="paper-plane" href="/agents/outbound-messaging">
    Multi-channel delivery pipeline for agent-initiated messages.
  </Card>

  <Card title="Cron & Scheduled Jobs" icon="clock" href="/agents/cron-jobs">
    Scheduled jobs, nudges, and the minion pattern.
  </Card>

  <Card title="Minion Pattern" icon="hand-holding-hand" href="/agents/minion-pattern">
    Inter-session task handoff for autonomous work.
  </Card>
</CardGroup>

## Agent Configuration

Agents are configured in `argent.json` under the `agents` key:

```json theme={null}
{
  "agents": {
    "defaults": {
      "model": "claude-sonnet-4-20250514",
      "maxTokens": 16384,
      "modelRouter": {
        "enabled": true
      }
    },
    "main": {
      "name": "Argent",
      "identity": "You are Argent, a personal AI assistant."
    }
  }
}
```

## Lifecycle

1. **Gateway starts** and loads agent configuration
2. **Channel message arrives** and is routed to the agent
3. **Agent processes** the message through the [agent loop](/agents/agent-loop)
4. **Response is sent** back through the originating channel
5. **Memory is updated** with relevant information from the interaction

The agent persists as long as the gateway is running. It wakes up on incoming messages and can also be triggered by the heartbeat system.