Skip to content
Qlinnt
All guides

Controlled example

Verify a Website Fix Without Mistaking Missing Coverage for a Pass

By Qlinnt · Updated

A finding disappearing from a report is not enough to prove a correction. In Qlinnt’s controlled browser tests, a fixture without an H1 produces a finding; a separately measured corrected fixture explicitly passes that check. Blocked or incomplete checks remain unverified. This is a test scenario, not a customer success story.

Two website documents linked together, with an amber uncertainty symbol and a mint check under a violet lens.
Editorial illustration: verify the changed condition with a new measurement.

Start with a controlled, deliberately incomplete page

The browser integration suite supplies its own HTML, public network responses and in-memory persistence double. It does not scan a customer site. The fixture has content and a canonical URL; the issue variant intentionally omits its main heading.

The target address in the test is intercepted by the browser and controlled fetch doubles. Do not copy it into a live scan; run the test suite to reproduce the scenario.

Simplified excerpt of the issue fixture
<main>
  <p>Real fixture content for evaluation.</p>
  <!-- Missing the main heading in the issue variant. -->
</main>

Change the measured condition, then measure again

The corrected fixture adds an H1. The integration assertions require seo.h1_count to be absent from findings and explicitly recorded as passed. This checks the completed outcome rather than assuming that every missing finding is a pass.

Each scenario is a separate controlled run. The comparison unit tests exercise how previous and current records are classified; this example does not claim a customer deployment or Google indexing improvement.

Simplified excerpt of the corrected fixture
<main>
  <h1>Fixture business</h1>
  <p>Real fixture content for evaluation.</p>
</main>

Keep an unavailable check distinct from a clean result

Another browser scenario triggers a POST and simulates failed public-file probes. The read-only guard blocks the write, and the exposed-file check remains inconclusive. The assertions do not count that uncertainty as passed.

Production sites can behave differently from fixtures. This exercise verifies a narrow implementation contract; it is not a measurement of overall accuracy, security or accessibility compliance.

Reproduce and interpret the assertions

In a project checkout with app and worker dependencies installed and a compatible Chromium available, run the controlled browser tests and comparison tests. CI requires a browser rather than silently skipping these checks.

After correcting your own website, deploy it and start a fresh scan. Read the check outcome and scope. Use Search Console separately to investigate indexing or rankings.

Project checkout; Playwright 1.50 uses the supported Ubuntu 22.04 CI runner
npm ci
npm ci --prefix worker
cd worker && npx playwright install --with-deps chromium
cd ..
npm test -- worker/src/scanner.integration.test.js src/lib/scan-comparison.test.ts

Does adding an H1 guarantee higher rankings?

No. This example verifies a heading check. Search usefulness, content, indexing and rankings require their own evidence.

Does a passed check mean the whole site is safe?

No. It describes the completed check and measured scope, not every risk or user journey.

Sources and further reading

Check your own public website

Start with a free scan. Verify your email to read every finding explained. Evidence and fix prompts are not included in the free diagnosis.

Scan my website