Write and Simulate Your Custom AI Agent’s Instructions

Instructions Prompt

Turn my agent brief into version 1 of operating instructions for my AI agent. Organize the instructions into these sections:
• Role and primary goal
• User and important context
• Process the agent should generally follow
• Decision rules
• Boundaries and actions requiring human approval
• Required output
• Definition of done
• What to do when information is missing
• Uncertain or conflicting
• Use clear, direct instructions
Don't add tools, models, connectors, memory, or technical implementation details that I haven't chosen yet. Don't expand the agent beyond the job defined in my brief. If you need to make an assumption, list it separately instead of silently adding it to the instructions. Here is my current agent brief.
[paste agent brief]

Simulation Prompt

For this exercise, simulate an AI agent that follows the instructions below. You don't have to access any tools, connectors, files, or external services unless I provide the information directly in this conversation. Follow the instructions as closely as possible, and don't pretend to take actions you can't actually perform. Here are the agent instructions:
[paste agent instructions]