scoutqa-test_skill

Overview

This skill runs AI-powered exploratory QA on web applications using the ScoutQA CLI. It is designed to start automated smoke tests, accessibility audits, user-flow verifications, and parallel test runs proactively after features are implemented. Use it to capture an execution ID and live browser URL immediately so you can monitor progress while tests continue remotely.

How this skill works

Compose a natural-language test prompt describing what to verify (goals and edge cases, not step-by-step actions) and run scoutqa with --url and --prompt. When invoking the Bash tool, set timeout: 5000ms to capture the ScoutQA execution ID and browser URL in the initial output; the test then continues running remotely and locally in the background. Share the execution ID and live URL with users, then fetch or review results later via the browser link or CLI.

When to use it

• After implementing a new feature (login, registration, checkout) to verify it works
• When a user explicitly asks to test a site, find bugs, or run exploratory QA
• For post-deployment smoke tests to validate critical flows
• To run accessibility audits (WCAG checks, keyboard navigation, screen readers)
• When you want parallel coverage: run multiple scoutqa invocations in one message

Best practices

• Write prompts that state what to explore and verify, not exact click sequences
• Always set the Bash tool timeout parameter to 5000ms to capture execution details
• Run multiple tests in parallel for different areas (auth, core features, accessibility)
• Provide credentials or extra context via scoutqa send-message if the agent gets stuck
• Report execution ID and live URL immediately so stakeholders can monitor progress

Example use cases

• Proactively test a newly merged registration form at http://localhost:3000/register
• Run an accessibility audit for a public landing page to find WCAG issues
• Start parallel runs: authentication, dashboard workflows, and mobile responsiveness
• Verify a checkout flow after changes to pricing or cart logic
• After a bug fix, run a targeted test to ensure the issue is resolved and no regressions occurred