Agents

An agent is the core unit in Clawployees. It's an AI-powered assistant that connects to communication platforms and handles conversations autonomously.

How Agents Work

Each agent runs as an isolated process with its own:

• Memory — conversation history and context
• Skills — tools and capabilities the agent can use
• Knowledge — documents and data the agent can reference
• Guardrails — rules about what the agent can and can't do

Agent Lifecycle

Created → Configured → Connected → Active → (Paused) → Archived

Routing

When a message arrives, the agent:

1. Checks if it should respond (channel rules, mention detection)
2. Loads relevant context from memory and knowledge base
3. Selects appropriate skills for the task
4. Generates and sends a response
5. Stores the interaction in memory

Memory

Agents maintain both short-term (conversation) and long-term (cross-conversation) memory. This allows them to remember user preferences and past interactions.

// Access agent memory programmatically
const memory = await agent.memory.search('user preferences');

Multi-Agent Collaboration

Agents can escalate to other agents or human operators when they encounter tasks outside their expertise. Configure escalation rules in your team settings.