Test a round-up rule actually collecting a round-up

The flow under test

The first three steps are configuration and the fourth is the only one that tests behaviour.

• The starting state - a zero balance and a badge confirming the rule is off.
• The rule is enabled - the badge flips to on.
• The multiplier is set - two times round-up, reflected in a current multiplier field.
• A purchase funds the vault - a balance of $0.80 with an entry naming the merchant.

The $0.80 figure is the whole test. A $24.60 purchase rounds up by $0.40, and the two times multiplier doubles it. That single number proves the rule fired, the multiplier applied and the vault was credited.

The test file

The final step leaves the vault page, drives a purchase on the card page, and returns. That round trip is what makes the assertion about behaviour rather than configuration.

---
mode: testing
url: https://my-testing-repo-main.vercel.app/revolut-clone-app/vault?reset=true
max_steps: 45
tags: [revolut, fintech, savings]
---

# Revolat 19.4: Savings vault

Catalog objective: enable the round-up vault and verify the rule is active.
Key assertion: the vault shows the round-up toggle on.

## Verify the vault starts off
Verify the "Rainy day" vault shows a balance of "$0.00" with a badge saying the round-up rule is off.

## Turn the round-up rule on
Click the toggle button and verify the badge now reads that the round-up rule is on.

## Choose a multiplier
Choose "2× round-up" and verify "Current multiplier" reads "2×".

## Verify the rule actually collects a round-up
Click the link through to the card page, click "Simulate purchase", then return to https://my-testing-repo-main.vercel.app/revolut-clone-app/vault and verify the vault balance reads "$0.80" with a round-up entry from "Metro Grocer".

The round-up entry is asserted by merchant name as well as amount, so a credit that appears without an attributable source fails.

Step by step coverage

Three configuration steps and one proof.

Without the fourth step this is a settings test. With it, the file verifies the feature customers are actually buying.

Running it

kane-cli testmd run tests/revolut/savings-vault_test.md

Add a variant at the default multiplier, since a doubled round-up and a single one differ by exactly the amount a bug would produce.

Background rules are the hardest thing to verify by hand and the easiest to leave uncovered. Kane CLI as a QA agent covers building them into a suite, and the three modes covers iterating on a multi-page flow interactively.