GitHub Repository: scoutqa-dot-ai/claude-code
What is Vibe Testing with Claude Code?
Vibe testing is a new approach to software testing where you describe what to test in plain English, and AI handles the execution. With the Scout skill installed in Claude Code, you can:- Ask Claude to “test the login flow” and get real results
- Run accessibility audits by saying “check WCAG compliance”
- Validate user journeys without writing a single test script
Features
Natural Language Testing
Describe tests conversationally. “Test the checkout flow” or “Find accessibility issues on the homepage.”
Accessibility Audits
Automatically identify WCAG compliance issues and accessibility barriers.
User Flow Validation
Verify critical journeys like signup, login, checkout, and onboarding work correctly.
Usability Analysis
Detect UX friction, confusing interfaces, and potential user frustrations.
Installation
Prerequisites
First, install the Scout CLI:Add the Skill to Claude Code
Usage
Once installed, simply ask Claude to run tests:- Login Flow
- Accessibility
- Smoke Test
- Checkout
Example Workflow
Troubleshooting
| Issue | Solution |
|---|---|
| Skill not found | Run claude plugin list to verify installation |
| Authentication errors | Run scoutqa auth login to re-authenticate |
| Tests failing to start | Verify target URL is publicly accessible |
Why Use Vibe Testing with Claude Code?
| Traditional Testing | Vibe Testing with Scout |
|---|---|
| Write and maintain test scripts | Describe tests in natural language |
| Brittle selectors break often | AI adapts to UI changes |
| Limited to predefined scenarios | Exploratory - finds unexpected issues |
| Requires testing expertise | Anyone can write tests |
Frequently Asked Questions
What is a Claude Code skill?
What is a Claude Code skill?
A Claude Code skill is a plugin that extends Claude’s capabilities. The Scout skill adds vibe testing functionality, allowing Claude to run real tests on your web applications when you ask.
How does vibe testing work in Claude Code?
How does vibe testing work in Claude Code?
When you ask Claude to test something, the Scout skill executes AI-powered exploratory tests on your target URL. Scout’s agent navigates your app, performs actions, checks for issues, and reports findings back to Claude - who then summarizes results for you.
Can I use this for accessibility testing?
Can I use this for accessibility testing?
Yes! Scout performs WCAG accessibility audits. Just ask Claude to “check accessibility” or “audit for WCAG compliance” on any URL.
Does this work with localhost?
Does this work with localhost?
Scout requires publicly accessible URLs. For localhost testing, use a tunneling service like ngrok to expose your local server, then test using the tunnel URL.
What's the difference between this and the Scout CLI?
What's the difference between this and the Scout CLI?
The Scout CLI is a standalone terminal tool. The Claude Code skill integrates Scout directly into your Claude Code workflow, so you can test without leaving your conversation with Claude.
Related
- Scout CLI - Run Scout from any terminal
- Cursor Rules - Use Scout in Cursor IDE
- Kiro Power - Use Scout in Kiro IDE
Testing localhost applications
Testing localhost applications
Scout requires a publicly accessible URL to run tests. Deployed environments like staging and production work immediately.For localhost testing, use a tunneling service like ngrok:Then ask Claude to test using the ngrok URL:*Third-party tunneling services are not affiliated with Scout. Use at your own risk.