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

# Agent skills

> Build with Sqwish using your coding agent.

Sqwish agent skills will give your coding agent context on D1’s question types,
request formats, and integration patterns.

## Installation

<Note>
  The Sqwish agent skill is not published yet. Setup and update instructions
  will be added here when it is available.
</Note>

## Example prompts

For now, share the [quickstart](/quickstart) and [API behavior](/api) pages with
your agent, then adapt one of these prompts to your project.

### Find a use case

```text theme={null}
Using the Sqwish documentation I shared, inspect this project and propose one
place where a binary, choice, or score question would help. Show the request
shape and how the application would use the answer.
```

### Build an integration

```text theme={null}
Add a server-side D1 integration following the Sqwish documentation I shared.
Read D1_API_URL and D1_API_KEY from environment variables. Preserve question
IDs, and reuse the same Idempotency-Key and request body after a timeout.
Include handling for request_in_progress and terminal failures.
```

## Next steps

Use [Route a support request](/examples/customer-support) as a starting example.
Review the questions and how your application uses each answer. The
[retry guide](/examples/reliable-retries) covers recovering interrupted requests;
[Create a prediction](/api-reference/predictions/create-a-prediction) documents
the complete request and response.