testing-expert

---

name: testing-expert description: Test quality inspection framework for reviewing test coverage, identifying gaps, and ensuring comprehensive validation. Trigger: test quality, coverage gaps, test inspection, assertion quality. disable-model-invocation: true

Example Test Inspection Report

Scenario: User Registration Feature

Engineer's Test Suite

# test_user_registration.py

def test_user_creation():
    """Test user creation"""
    user = create_user("[e-mail]", "password123")
    assert user

def test_login():
    """Test login"""
    user = create_user("[e-mail]", "password123")
    result = login("[e-mail]", "password123")
    assert result

def test_duplicate_email():
    """Test duplicate email"""
    create_user("[e-mail]", "password123")
    user2 = create_user("[e-mail]", "password456")
    assert user2

---

QA Inspection Report

Test 1: test_user_creation()

Claimed Intent: Test user creation Actually Tests: Object instantiation only

Issues Found

🔴 CRITICAL: Worthless Assertion

assert user  # Line 5

• Would pass with corrupt data
• Would pass with wrong email
• Would pass with invalid password hash
• Would pass if database save failed

Mental Debug: User object with garbage data would pass this test.

🟠 HIGH: Missing Validation Tests

• No test for invalid email format
• No test for weak password
• No test for duplicate email (separate test exists but broken)
• No test for SQL injection in email

🟠 HIGH: No Persistence Verification

• Test doesn't verify user is saved to database
• Test doesn't verify user can be retrieved
• Transaction might be rolled back

🟡 MEDIUM: Vague Test Name

• "test_user_creation" doesn't describe what should happen
• Better: "test_user_creation_with_valid_data_persists_to_database"

Suggested Improvements

def test_user_creation_with_valid_data_persists_to_database():
    """Creating a user with valid email and password should:
    1. Create user object with correct attributes
    2. Save to database
    3. Hash password (not store plaintext)
    4. Set default role and active status
    """
    email = "[e-mail]"
    password = "SecurePass123!"

    user = create_user(email, password)

    # Verify user object
    assert user.id is not None, "User should have an ID after creation"
    assert user.email == email, "Email should match input"
    assert user.is_active is True, "New users should be active by default"
    assert user.role == "user", "New users should have 'user' role"
    assert user.created_at is not None, "Created timestamp should be set"

    # Verify password is hashed, not plaintext
    assert user.password_hash != password, "Password should be hashed"
    assert len(user.password_hash) > 50, "Password hash should be substantial"

    # Verify persistence
    retrieved_user = User.get_by_email(email)
    assert retrieved_user is not None, "User should be retrievable from database"
    assert retrieved_user.id == user.id, "Retrieved user should match created user"

def test_user_creation_with_invalid_email_format_raises_validation_error():
    """Creating a user with malformed email should raise ValidationError"""
    invalid_emails = [
        "not-an-email",
        "@example.com",
        "test@",
        "test [e-mail]",
        "",
    ]

    for invalid_email in invalid_emails:
        with pytest.raises(ValidationError) as exc:
            create_user(invalid_email, "password123")

        assert "email" in str(exc.value).lower()
        assert "invalid" in str(exc.value).lower()

def test_user_creation_with_weak_password_raises_validation_error():
    """Creating a user with weak password should raise ValidationError"""
    weak_passwords = [
        "123",          # Too short
        "password",     # No numbers
        "12345678",     # No letters
        "",             # Empty
    ]

    for weak_password in weak_passwords:
        with pytest.raises(ValidationError) as exc:
            create_user("[e-mail]", weak_password)

        assert "password" in str(exc.value).lower()

Risk Level: 🔴 CRITICAL Action: ❌ BLOCK - Core functionality not tested Estimated Fix Time: 30 minutes

---

Test 2: test_login()

Claimed Intent: Test login Actually Tests: Function call completes

Issues Found

🔴 CRITICAL: Worthless Assertion

assert result  # Line 11

• Passes with any truthy value
• Doesn't verify session/token
• Doesn't verify user authentication state

🔴 CRITICAL: Missing Negative Tests

• No test for wrong password
• No test for non-existent user
• No test for locked account
• No test for expired credentials

🟠 HIGH: No Session Verification

• Doesn't verify authentication token
• Doesn't verify session expiry
• Doesn't verify user context in session

🟡 MEDIUM: Test Depends on Previous Test

• Creates user in this test
• Should use fixture or setup
• Tests should be independent

Suggested Improvements

@pytest.fixture
def registered_user():
    """Fixture providing a registered user for login tests"""
    user = create_user("[e-mail]", "SecurePass123!")
    yield user
    # Cleanup if needed
    User.delete(user.id)

def test_login_with_valid_credentials_returns_authenticated_session(registered_user):
    """Logging in with correct email and password should:
    1. Return authentication token/session
    2. Set authenticated state
    3. Include user context
    4. Set appropriate expiry
    """
    session = login(registered_user.email, "SecurePass123!")

    assert session is not None, "Login should return session"
    assert session.is_authenticated is True, "Session should be authenticated"
    assert session.user_id == registered_user.id, "Session should contain user ID"
    assert session.token is not None, "Session should have authentication token"
    assert session.expires_at > datetime.now(), "Session should have future expiry"
    assert (session.expires_at - datetime.now()).seconds >= 3600, "Session should last at least 1 hour"

def test_login_with_wrong_password_raises_authentication_error(registered_user):
    """Logging in with incorrect password should raise AuthenticationError"""
    with pytest.raises(AuthenticationError) as exc:
        login(registered_user.email, "WrongPassword")

    assert "Invalid credentials" in str(exc.value)
    assert "password" in str(exc.value).lower()

def test_login_with_nonexistent_email_raises_authentication_error():
    """Logging in with non-existent email should raise AuthenticationError"""
    with pytest.raises(AuthenticationError) as exc:
        login("[e-mail]", "password")

    assert "Invalid credentials" in str(exc.value)
    # Note: Don't reveal if email exists (security)

def test_login_with_locked_account_raises_account_locked_error(registered_user):
    """Logging in to locked account should raise AccountLockedError"""
    lock_account(registered_user.id)

    with pytest.raises(AccountLockedError) as exc:
        login(registered_user.email, "SecurePass123!")

    assert registered_user.email in str(exc.value)

def test_login_with_empty_password_raises_validation_error(registered_user):
    """Logging in with empty password should raise ValidationError"""
    with pytest.raises(ValidationError) as exc:
        login(registered_user.email, "")

    assert "password" in str(exc.value).lower()
    assert "required" in str(exc.value).lower()

Risk Level: 🔴 CRITICAL Action: ❌ BLOCK - Authentication not actually tested Estimated Fix Time: 45 minutes

---

Test 3: test_duplicate_email()

Claimed Intent: Test duplicate email handling Actually Tests: Second user creation succeeds (WRONG!)

Issues Found

🔴 CRITICAL: Test is Backwards

user2 = create_user("[e-mail]", "password456")
assert user2  # Line 17

• This test expects duplicate creation to SUCCEED
• It should expect it to FAIL with an error
• Test passes when it should fail
• This is testing the opposite of what's needed

🔴 CRITICAL: False Confidence

• Production bug: duplicate emails are allowed
• Test claims to verify duplicate prevention
• Test actually verifies duplicates work
• QA might approve thinking it's covered

🟡 MEDIUM: Same Email Issue as Other Tests

• If this fixed to expect error, needs all improvements from Test 1

Suggested Fix

def test_create_user_with_duplicate_email_raises_integrity_error():
    """Creating a user with an email that already exists should:
    1. Raise IntegrityError or ValidationError
    2. Not create duplicate user in database
    3. Preserve existing user data
    """
    email = "[e-mail]"

    # Create first user
    user1 = create_user(email, "FirstPassword123!")
    initial_count = User.count()

    # Attempt to create duplicate
    with pytest.raises((IntegrityError, ValidationError)) as exc:
        create_user(email, "SecondPassword456!")

    assert "email" in str(exc.value).lower()
    assert "duplicate" in str(exc.value).lower() or "exists" in str(exc.value).lower()

    # Verify no new user created
    assert User.count() == initial_count, "User count should not increase"

    # Verify original user unchanged
    original_user = User.get_by_email(email)
    assert original_user.id == user1.id, "Original user should be intact"
    assert original_user.verify_password("FirstPassword123!"), "Original password should work"
    assert not original_user.verify_password("SecondPassword456!"), "New password should not work"

Risk Level: 🔴 CRITICAL Action: ❌ BLOCK - Test verifies opposite of requirement Estimated Fix Time: 20 minutes

---

Summary Report

Overall Assessment

Test Suite Quality: 🔴 FAILING

Critical Issues: 3

• Test 1: Doesn't actually test user creation
• Test 2: Doesn't actually test authentication
• Test 3: Tests opposite of requirement

Total Tests: 3 Effective Tests: 0 Coverage: High (claims) Protection: None (reality)

Risk Assessment

Production Risk: 🔴 EXTREME

Current test suite provides zero protection against:

• Data corruption in user creation
• Authentication bypass
• Duplicate email registration
• Password security issues
• Database integrity issues

Confidence Level: 0% - Tests passing means nothing

Required Actions

Immediate (Block Merge)

1. Rewrite all three tests with proper assertions
2. Add negative test cases (12+ tests needed)
3. Verify tests catch intentional bugs
4. Add fixture for test user management

Follow-up (Required for completion)

1. Add edge case tests (15+ additional tests)
2. Add integration tests for full registration flow
3. Add security tests (SQL injection, XSS, etc.)
4. Add performance tests for registration endpoint

Estimated Timeline

• Fix critical issues: 2-3 hours
• Complete test suite: 1 day
• Review and iteration: 0.5 days

Total: 1.5-2 days for proper test coverage

Recommendation

❌ BLOCK MERGE

Do not approve this PR. Tests provide false confidence and mask critical bugs.

Evidence:

• All tests would pass with completely broken functionality
• Duplicate email test verifies the opposite of requirements
• No actual behavior is verified

Next Steps:

1. Engineer rewrites tests following examples above
2. QA re-inspects rewritten tests
3. QA verifies tests catch intentional bugs
4. Only then approve merge

---

Lessons for Engineer

What Went Wrong

1. Wrote tests after code - Led to tests that just confirm code runs
2. Weak assertions - "assert x" proves nothing
3. No mental debugging - Didn't verify tests catch bugs
4. No negative testing - Only tested happy path
5. Misunderstood duplicate test - Test verified opposite

How to Improve

1. Write tests first (TDD) - Prevents these issues
2. Specific assertions - Verify exact values
3. Mental debugging - Break code, ensure test fails
4. Test failures explicitly - Every success needs failure test
5. Read test name carefully - Test what you claim to test

TDD Would Have Prevented This

If tests were written first:

# Write this FIRST (it will fail):
def test_user_creation_with_valid_data_persists_to_database():
    user = create_user("[e-mail]", "password")
    assert user.email == "[e-mail]"  # Will fail until create_user works
    ...

# Then implement create_user to make it pass

See the Test-Driven Development skill for complete TDD workflow (available in the skill library for comprehensive TDD guidance).

---

Sign-off

QA Inspector: [Your name] Date: [Date] Status: ❌ REJECTED Reason: Tests provide zero protection, must be rewritten Re-inspection Required: Yes

---

This is what thorough test inspection looks like. Better to catch these issues now than in production.