Core Docs - Core Concepts - Configure an Agent

Rock Version: v20.0
Last Modified: 2026-10-01 11:28 AM

This is where an agent takes shape. You give it a name, a purpose and the boundaries that keep it on track. Agents are configured under Admin Tools > Settings > AI Agents.

Before configuring your first agent, set your organization context under Settings on the same page. That context applies to every agent you build, so it's worth doing once up front. See Setting Your Organization Context for details.

Agent Settings

To add a new Agent, select ti ti-robot Agents then click ti ti-square-plus.

Name - Give the agent a name that says what it's for. "Staff Chat Assistant" tells your team far more than "Agent 1."

Type - Choose Chat or MCP. You can't change this after you save, so decide up front. See the Type section.

Audience - Set this to Internal or Public. It shapes how the agent's tools behave. See the Audience section.

Description - Record which team or purpose the agent serves. This is for your notes only and doesn't change how the agent behaves.

Instructions - Tell the agent how to act. See the Instructions section.

Avatar - Give the agent a face in the chat panel by uploading an image. This is optional.

Model - Pick how much power the agent brings to each request. Medium is the default. It's fast, costs less and handles everyday chat and lookups well. High uses a more capable model built for heavier reasoning, so save it for agents that take on complex, multi-step work. For most agents, leave this at Medium.

Reasoning Effort - Decide how much thinking the agent does before it answers, from Minimal up to Extra High. Higher settings give more thorough answers but take longer and use more tokens. Low is the default and suits quick, conversational requests. Move up to High for an agent that handles complex, multi-step work.

Avatar, Model and Reasoning Effort appear only on Chat agents. An MCP agent asks for a Slug instead, the name that identifies the agent in its connection URL, and its Advanced Settings hold only Exclude System Skills. See Set Up an MCP Agent.

Advanced Settings

Most organizations never need to change these settings.

Auto Summarize Threshold - Keep long conversations focused. Once a conversation reaches this many tokens, the agent summarizes earlier messages. Leave it blank and Rock uses 60,000.

Exclude System Skills - Remove Rock's built-in system skills from this agent. They handle core functionality, so leave this unchecked unless you're sure. Off by default.

Current Person Template - Control what the agent knows about the person who's logged in, using a Lava template. Leave it blank and Rock uses a default template, which is right for almost everyone.

Type

Chat agents live inside Rock. Staff access them through the docked chat panel, which you can access from anywhere in Rock by clicking the ti ti-sparkles agent button. This is the right starting point for most organizations.

MCP agents connect Rock to an external AI tool running on someone's computer, like Claude Desktop. Once connected, that tool can look up records and take actions in Rock using the person's own security permissions. See Set Up an MCP Agent for the full walkthrough.

Audience

Audience is one of the most important decisions you'll make for each agent.

Instructions

Instructions are where you shape how the agent behaves. Think of it as a short onboarding note: tell the agent how to act, what to prioritize and what to avoid.

A few things worth including:

Keep it short. A long list of rules is harder to follow consistently than a clear set of priorities. Start simple and add only when you notice a real gap. Only add what's needed.

Testing

Before sharing an agent with your team, test it yourself, more than once. AI doesn't produce identical results every run, so test your most important prompts at least a few times and look for consistency.

Also test as the people who will use it. Log in as a volunteer and try it. If the agent surfaces something that person shouldn't see, the fix is in your Rock security, not the agent configuration.