I need to stay on topic and focus on answering questions about N8N Docs. Let me help you with information about how agents are implemented in n8n.

Agents in n8n

In n8n, an agent is a decision-making entity that uses language models to determine which actions to take, unlike chains that follow a predetermined sequence of operations.

How Agents Work in n8n

n8n provides an AI Agent node that can function in various capacities depending on the settings you choose. The Agent node offers six different agent types:

1. Tools Agent (default) - Recommended for most use cases, this agent uses external tools and APIs to perform actions and retrieve information.

2. Conversational Agent - Designed for human-like conversations, maintaining context and understanding user intent.

3. OpenAI Functions Agent - Works specifically with OpenAI functions models.

4. Plan and Execute Agent - Creates a high-level plan and executes it step by step.

5. ReAct Agent - Reasons about tasks, determines necessary actions, and executes them in cycles.

6. SQL Agent - Uses SQL databases as data sources, converting natural language to SQL queries.

When you execute a workflow containing an agent, the agent runs multiple times. For example, it may perform an initial setup, followed by a run to call a tool, then another run to evaluate the tool response before responding to the user.

Each agent type has specific configuration options and capabilities that allow it to interact with different parts of your workflow.