Test search with size and price filters

The checks that matter

A filter is a claim about every result it leaves behind, so the only honest test opens one of those results and checks the claim held.

The last row is the assertion that matters. A filter that returns a product whose size 8 is out of stock has technically filtered and practically lied, and only a check on the product page catches it.

The same flow as a test file

Each filter gets its own step so the result count can be asserted between them. Combining them into one instruction would hide which of the two narrowed the set incorrectly.

---
mode: testing
url: https://my-testing-repo-main.vercel.app/flipkart-clone-app?reset=true
max_steps: 45
tags: [flipkart, e-commerce, search]
---

# Flipkort 3.1: Search with filters

Catalog objective: search for running shoes, filter by size and price band, then open a result.
Key assertion: the result respects both the size and the price filter.

## Search for running shoes
Type "running shoes" into the search box, click "Search", and verify the results page shows "4 results".

## Filter by size 8
Click the "8" size filter and verify the result count line contains "size 8".

## Narrow by price band
Click the "₹1,500 - ₹3,000" price filter and verify the results show "1 result" and list "Trailburst Running Shoes".

## Open the filtered result
Click "Trailburst Running Shoes" and verify the product page shows the price "₹2,499.00" and that size "8" can be selected.

## Confirm the size is genuinely available
Click size "8" and verify a badge appears reading "Size 8 · 12 left".

Run it with:

kane-cli testmd run tests/flipkart/search-with-filters_test.md

Run output

Five passes across 267 seconds, the longest run in the storefront set. Opening the filtered result was the slowest step at 66.9 seconds, since it loads a full product detail page and asserts two separate things on it.

---
test: ../search-with-filters_test.md
status: passed
started: 2026-09-15T09:37:30.270Z
duration_s: 267
session_id: 88afe998-569d-44a9-b338-2438c0668fba
---

# Flipkort 3.1: Search with filters - Result

## Search for running shoes ✓ passed (52.3s)
md5: 829741b25f9c4c5f47a9d88aaf2cb1db
Type "running shoes" into the search box, click "Search", and verify the results page shows "4 results".

## Filter by size 8 ✓ passed (47.4s)
md5: 85c7498ae7206c78719ef1aadf3d1969
Click the "8" size filter and verify the result count line contains "size 8".

## Narrow by price band ✓ passed (46.1s)
md5: 5f9a5f3a4b9eb1bec636527abe782f53
Click the "₹1,500 - ₹3,000" price filter and verify the results show "1 result" and list "Trailburst Running Shoes".

## Open the filtered result ✓ passed (66.9s)
md5: e61b486005c91977658793bffc167f3d
Click "Trailburst Running Shoes" and verify the product page shows the price "₹2,499.00" and that size "8" can be selected.

## Confirm the size is genuinely available ✓ passed (51.7s)
md5: c8e870c151a2095b906b0b34c1305555
Click size "8" and verify a badge appears reading "Size 8 · 12 left".

Taking it further

The result is a search test that checks the promise a filter makes rather than the number it displays. It is the difference between knowing the facet component runs and knowing the catalogue behind it is consistent.

Catalogues change more often than test code, which makes generated coverage worth considering. From a PRD to evidence covers building a requirement-linked suite, and browser agents is useful background on how the page is interpreted at run time.