> ## Documentation Index
> Fetch the complete documentation index at: https://docs.asteragents.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Agent Prompt Guide

> Learn how to effectively communicate with AI agents

## Introduction

A prompt is the input or instruction you provide to an AI agent. The agent will respond based on:

* its instructions (part of agent configuration by your Admins)
* its understanding of your prompt

<Note>
  Think of agents like your co-workers:

  * They want to be helpful, but…
  * They can't read your mind and, most importantly…
  * They tend to take things very literally
</Note>

## Important Terms

<AccordionGroup>
  <Accordion icon="message" title="Prompt">
    The message you (the user) send to the agent
  </Accordion>

  <Accordion icon="robot" title="Agent">
    An individual AI bot with a primary objective and access to tools to help it accomplish that objective
  </Accordion>

  <Accordion icon="wrench" title="Tools">
    Agents can call tools to make API calls to other apps or systems (i.e. lookup information in a database or send an e-mail)
  </Accordion>

  <Accordion icon="code" title="Tool Call">
    A single invocation of a tool, including parameters that are passed to that tool (i.e. for sending an email, the agent would choose the send\_email tool and also send a parameter containing the email content)
  </Accordion>

  <Accordion icon="brain" title="Model">
    The specific AI model that an agent uses (the model is like the agent's brain), such as gpt-4 from OpenAI or Claude 3.5 Sonnet from Anthropic
  </Accordion>

  <Accordion icon="clipboard-check" title="Evals">
    Tests that are used to evaluate your agent, with an input prompt and an expected result
  </Accordion>

  <Accordion icon="chart-network" title="Embeddings">
    A mathematical representation of text as a series of numbers. Allows for computation such as "how similar is each chunk of text to my search term"
  </Accordion>
</AccordionGroup>

## Tips for Writing Great Prompts

<CardGroup>
  <Card title="Provide Rich Context" icon="circle-info">
    The more context you can share, the better. The agent can only be as helpful as the instructions you give it.

    **Example:**
    Instead of saying "Find information on software version upgrade", say "I'm helping a customer upgrade their version of \[software name] because they finally have internal bandwidth to do the upgrade. Help me share the right instructions and links with them to ensure their upgrade goes smoothly."
  </Card>

  <Card title="Focus on Positive Instructions" icon="circle-check">
    Don't tell AI what NOT to do. Instead, be specific and targeted with your instructions of what TO do.
  </Card>

  <Card title="Share Your Objective" icon="bullseye">
    Agents can make multiple tool calls and adjust based on findings, but they need to know your overall goal, not just the first step.
  </Card>

  <Card title="Explore Capabilities" icon="compass">
    If you're not sure what your agent can do, just ask! The agent can summarize its tools and suggest productive ways to use them.
  </Card>
</CardGroup>

### Additional Best Practices

<AccordionGroup>
  <Accordion icon="comments" title="Have a Conversation">
    If the agent doesn't succeed at first, keep trying! Have a back and forth conversation and encourage the agent to try different approaches.
  </Accordion>

  <Accordion icon="markdown" title="Use Markdown">
    Agents are trained to understand Markdown formatting. Use it for structured content like lists and headers.
  </Accordion>

  <Accordion icon="magnifying-glass" title="Ask for Citations">
    Ask the agent to explain its chain of thought and include specific citations when you want to understand its reasoning.
  </Accordion>
</AccordionGroup>