THE BASIC ARCHITECTURE OF AN AGENT

Before configuring an Agent on F68 AI, describe the job at the design level.

OBJECTIVE
What job does the Agent exist to solve?

↓

INPUT
What information does the Agent receive?

↓

SYSTEM PROMPT
Which principles and ways of working must be followed?

↓

CONTEXT
What does the Agent need to know to understand the task correctly?

↓

TOOL
Which data or capabilities does the Agent need access to?

↓

PROCESS
Which steps must happen before the result is produced?

↓

OUTPUT
What structure must the final result have?

↓

VERIFICATION
How do you check that the result is good enough to use?

This is a mental framework for designing an Agent, not eight separate configuration fields in the interface.

When deploying:

• principles and process can live in the System Prompt;
• Knowledge/Data can live in the Context;
• access to the outside world lives in the Tool.