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

# Dialog-RSN-1 prompt cookbook

> Ready-to-adapt `instructions` text, grouped by behavior: waiting for a complete sequence, confirming before a destructive tool call, and shaping how a reply is spoken.

Every recipe here is `instructions` text you can paste into `session.update` and adapt to your own agent. None of it is a new field or event: it's the same `instructions` string described in the [prompt guide](/dialog-rsn-1/prompt-guide), applied to a specific behavior.

<CardGroup cols={2}>
  <Card title="Turn-taking" icon="waveform-lines" href="/dialog-rsn-1/prompt-cookbook/turn-taking">
    Wait for a complete sequence, offer a fallback when a caller's stuck, and react to sounds a transcript would miss.
  </Card>

  <Card title="Tool calling" icon="wrench" href="/dialog-rsn-1/prompt-cookbook/tool-calling">
    Confirm before anything destructive, narrate results instead of tool names, recover from a tool error, and search with citations.
  </Card>

  <Card title="Response style" icon="waveform" href="/dialog-rsn-1/prompt-cookbook/response-style">
    Say numbers the way a caller expects, keep replies short, and tag a reply with an emotion for your TTS.
  </Card>

  <Card title="Grounding" icon="crosshairs" href="/dialog-rsn-1/prompt-cookbook/grounding">
    Give ASR a known list of names and terms so it only has to get close.
  </Card>
</CardGroup>

For the reasoning behind a recipe before you adapt it, start with the [prompt guide](/dialog-rsn-1/prompt-guide) instead.