Write Test Plans That Catch Bugs Before Users Do — QA Engineering

By the end of this page, you will understand how QA Engineers write and execute test plans using pytest and vitest — and how AI agents can generate comprehensive test suites automatically.

Testing & Verification — The 2-Minute Overview

Think about the last time you bought a car. You didn't see the thousands of crash tests, brake tests, engine stress tests, and safety inspections behind that vehicle. You just turned the key and drove. But somebody had to verify that every part works, every edge case is handled, and every safety standard is met — before it reached the showroom. That verification is QA Engineering. The diagram below is that map, zoomed out.

You Already Know QA — You Just Don't Know It Yet

You've been doing QA every time you proofread an important email before hitting send. Let's prove it.

Imagine you're sending a job application email to your dream company:

✉️ The Email Proofreading Analogy

Step 1 — You check structure: Subject line present? Greeting correct? Attachment included?

🔗 QA Layer: ① UNIT TESTING — Test individual components in isolation. Does each function return the correct output?

Step 2 — You check flow: Does paragraph 1 connect to paragraph 2? Does the closing match the tone?

🔗 QA Layer: ② INTEGRATION TESTING — Test component interactions. Do modules work together correctly?

Step 3 — You check edge cases: What if they open it on mobile? What if the attachment is too large? What if the link is broken?

🔗 QA Layer: ③ EDGE-CASE TESTING — Test boundaries, errors, and unusual scenarios.

Step 4 — You read it as the recipient: Does it make sense to someone who doesn't know your context?

🔗 QA Layer: ④ ACCEPTANCE TESTING — Does the product meet the user's acceptance criteria?

The Complete Mapping

You just learned QA without writing a single test.

The 5 Pillars of QA Engineering

1. Test Strategy

Not all tests are equal. The strategy defines what to test, at what level, with what priority.

The testing pyramid: many unit tests (fast, cheap), fewer integration tests (moderate), few e2e tests (slow, expensive). The QA Engineer maps each acceptance criterion to a test level and sets coverage targets.

2. pytest (Backend Testing)

pytest is the Python testing standard — simple, powerful, and extensible.

Fixtures for setup/teardown. Parametrize for testing multiple inputs. Markers for categorizing tests. Coverage reports for identifying gaps.

3. vitest (Frontend Testing)

vitest is the Vite-native testing framework — fast, modern, and compatible with Jest.

Component testing for React components. Snapshot testing for UI regression. Mock testing for API calls. DOM testing for user interactions.

4. Definition of Done

Code isn't "done" when it compiles. It's "done" when all tests pass, coverage meets target, and acceptance criteria are verified.

The Definition of Done is a checklist: all unit tests pass, integration tests pass, coverage > 80%, no critical defects, acceptance criteria verified, code reviewed, and merged to DEV.

5. Defect Reporting & Triage

A bug report without reproduction steps is noise. A bug report with root cause is a gift.

Good defect reports include: summary, steps to reproduce, expected vs. actual behavior, severity, and root cause hypothesis. Triage categorizes by severity and assigns to the right developer.

The Complete Mapping

That's it. Master these 5 pillars, master QA.

Try It Yourself — A Starter Prompt for Test Plan Generation

This prompt gives you a working starting point. For the complete prompt — with test-to-AC mapping, coverage gap analysis, and defect triage workflows — see the full course chapter →.

You are a Senior QA Engineer with experience in pytest and vitest.

I need a test plan for:

{{PASTE YOUR FEATURE CODE OR REQUIREMENTS}}

Cover these 5 areas:

1. TEST STRATEGY — Map each requirement to a test level (unit, integration, e2e).
2. BACKEND TESTS — Write pytest test cases for every backend function. Include edge cases.
3. FRONTEND TESTS — Write vitest test cases for every UI component. Include user interaction tests.
4. COVERAGE TARGETS — Set coverage targets per module and identify potential gaps.
5. DEFINITION OF DONE — Define the checklist for "done" including test pass rates and coverage thresholds.

For each area, provide: the test cases and a brief justification.

Format as a structured document with tables where appropriate.

What This Prompt Covers vs. What It Misses

The Lite Prompt gets you to ~60% quality. Good enough to have a test plan. Not good enough to catch the bugs that reach production.

Real-World Example: Test Plan for a User Registration Feature

The Requirement

"Test a user registration feature: email validation, password strength check, duplicate account detection, welcome email trigger, and database record creation."

Lite Prompt Output — High-Level Test Plan

① Test Strategy

Unit tests for email validation and password check. Integration test for registration flow. E2E test for full signup journey.

② Backend Tests (pytest)

test_valid_email, test_invalid_email, test_strong_password, test_weak_password, test_duplicate_email, test_registration_creates_user.

③ Frontend Tests (vitest)

test_form_renders, test_submit_with_valid_data, test_error_on_invalid_email.

④ Coverage Target

80% line coverage for registration module.

⑤ Definition of Done

All tests pass. 80% coverage. No P0 bugs.

What a QA Lead Would Catch

The pattern: The Lite Prompt asks "what are the test cases?" The full course asks "what are the test cases, what do they miss, and what bug escapes to production?"

Ready to Write Test Plans That Catch Everything?

• ✅ The complete prompt with AC mapping, negative test generation, and flaky test detection
• ✅ An AI agent that writes and executes test plans automatically
• ✅ Assessment + coding challenges to verify you can test, not just describe tests

Enroll in the Fresh Graduate AI SDLC Course →

Go from "I write tests" to "I write test suites that catch bugs before users do."