Defining Custom Agent Scope

What Agent Scope Means

An agent’s scope defines:

• What the agent is responsible for

• What types of requests it should handle

• What requests it should decline or redirect

Clear scope helps users understand when to use the agent and helps the agent respond consistently.

One Agent, One Job

A custom agent should have one primary job.

A simple rule to define scope:

If you can describe when to use the agent in one sentence, the scope is likely correct.

Good examples:

• “Use this agent to summarize internal documents.”

• “Use this agent to answer customer support questions.”

If you need multiple sentences, or phrases like “and also”, the scope is usually too broad.

Define What Is Out of Scope

Good agents don’t try to answer everything.

A well-scoped agent should:

• Ignore requests outside its responsibility

• Ask users to switch agents when a request is out of scope

• Avoid guessing or filling in missing information

Explicit boundaries help keep responses accurate and predictable.

Common Scoping Mistakes

Avoid these common mistakes:

• Combining multiple unrelated tasks into one agent

• Writing instructions that are too broad or generic

• Expanding an agent instead of creating a new one

• Trying to replace the General Assistant with a single custom agent

When an agent produces inconsistent results, unclear scope is often the root cause.

Why Scope Matters

Clear scope helps:

• Improve response consistency

• Reduce incorrect or irrelevant answers

• Keep system prompts simpler

• Make agents easier to maintain over time

If a custom agent feels unreliable, reviewing its scope is usually the best first step.

What to Read Next

• Writing an Effective Custom Agent System Prompt

• Tools, Permissions, and Knowledge Access

• When to Create a Custom Agent (and When Not To)