---
title: "Sandbox and Testing"
description: "How to rehearse Coverage Cat's delegated flows safely before live traffic, with the published consumer and delegated umbrella sandbox paths."
canonical: "https://www.coveragecat.com/ai/sandbox-testing"
last-updated: "2026-11-11"
---

# Sandbox and Testing

Coverage Cat exposes sandbox-friendly paths so agents can test their control flow before sending real customers into live purchase workflows.

## Delegated umbrella sandbox

The public umbrella docs also describe sandbox sessions. Use them to validate draft and quote-loop behavior, but do not assume every post-quote step is supported in sandbox. Keep your test runner aligned with the published skill docs and OpenAPI notes.

## Testing advice

- Start from the discovery document so your runner uses real published endpoints.
- Use the HTML guides for sequence and guardrails, then pull the markdown skill file your runtime needs.
- Keep sandbox and live traffic clearly separated in logs, keys, and test data.

## Next steps

- **Learn:** Read [What Coverage Cat does](/what-we-do), browse [insurance guides](/insurance-types), or scan the [FAQ](/faq).
- **Estimate:** Use the [home insurance calculator](/tools/home), [umbrella insurance calculator](/tools/umbrella), or [carrier price comparison](/tools/auto-estimates).
- **Compare:** Start a [Coverage Cat quote comparison](https://www.coveragecat.com/intake) when you want licensed broker help and real quotes.
- **Integrate:** Explore the [Coverage Cat Agent API hub](/ai/skills), the [discovery document](/api/agent), and the [OpenAPI spec](/api/agent/openapi.yaml).

New homeowners application and quote workflows are unavailable and do not currently offer a sandbox. Existing homeowners policy support remains available.
