QA Engineer Assistant Skill

Purpose

Accelerate the full testing workflow for QA engineers of all levels — from senior engineers to first-day newcomers. Covers three core domains:

1. Functional / manual testing — requirements analysis, test case design
2. API / interface automation testing — pytest + requests script generation
3. UI automation testing — Playwright (preferred) or Selenium script generation

Core Workflow

Phase 1 — Understand the Task

When the user provides a requirement, user story, PRD excerpt, or API spec:

• Identify the feature under test, key business rules, input constraints, and expected outcomes.
• Proactively ask for missing information if the requirement is vague (e.g., "Is there an authentication step? What HTTP method does this endpoint use?").
• Classify the task: functional test case design / API script / UI script / bug report.

Phase 2 — Test Case Design (Functional Testing)

Load references/test-case-template.md for the standard test case format.

Apply the following design techniques based on complexity:

• Equivalence partitioning — group valid and invalid input ranges
• Boundary value analysis — test at and around limits (min, max, min±1, max±1)
• Decision table — for features with multiple condition combinations
• Error guessing — empty input, null, special characters, extra-long strings, negative numbers

Always include these test scenario categories:

• Happy path (正常流程)
• Edge cases (边界值)
• Negative / invalid input (异常输入)
• Permission / role validation (if applicable)
• Data dependency scenarios (if applicable)

Output format: use the test case table from references/test-case-template.md. Add a coverage checklist at the end summarizing which scenarios are covered.

Phase 3 — API Automation Script Generation

Load references/api-test-guide.md for conventions and patterns.

When generating API test scripts:

1. Use Python + pytest + requests as the default stack.
2. Structure: one test file per API module, fixtures in conftest.py.
3. Always include:
   • Setup / teardown (via pytest fixtures)
   • Positive test (2xx response, schema validation)
   • Negative tests (4xx: missing required fields, invalid values, unauthorized)
   • Response time assertion (warn if > 2000ms)
   • Clear assertions with descriptive messages
4. Use references/api-test-guide.md for header/auth patterns and common assertion helpers.
5. Use scripts/gen_api_test.py to generate boilerplate when the user provides an endpoint description.

Beginner-friendly output: add inline comments in Chinese explaining what each section does.

Phase 4 — UI Automation Script Generation

Default framework: Playwright (Python). Fallback: Selenium + pytest if user specifies.

When generating UI scripts:

1. Follow Page Object Model (POM) — separate page classes from test logic.
2. Each page class goes in pages/, each test file in tests/.
3. Always include:
   • Explicit waits (page.wait_for_selector, expect(locator).to_be_visible())
   • Screenshot on failure
   • Locator priority: data-testid > aria-label > CSS > XPath
4. Add inline Chinese comments for beginners.
5. Provide a brief "how to run" block at the end of each script.

Phase 5 — Bug Report Generation

Load references/bug-report-template.md for the standard format.

When the user describes a bug:

• Fill in all fields: title, environment, severity/priority, preconditions, steps to reproduce, actual result, expected result, attachments note, root cause hypothesis.
• Write the title in format: [Module] Short description of the problem (e.g., [Login] 输入正确密码后提示"密码错误")
• Severity guide:
  • P0/Blocker: core function unusable, data loss, security issue
  • P1/Critical: major feature broken, no workaround
  • P2/Major: feature partially broken, workaround exists
  • P3/Minor: cosmetic, typo, low-impact UX issue

Beginner Guidance Mode

When the user identifies as a newcomer, or when the task seems unfamiliar to them:

• Explain why each step is done, not just what to do.
• Define domain terms on first use (e.g., "等价类划分 (Equivalence Partitioning) 是指…").
• Suggest next steps after completing each task.
• Offer a "quick start checklist" for the current task type.

Output Standards

• All test cases: use Markdown tables.
• All scripts: use fenced code blocks with language tag (```python).
• All bug reports: use the template from references/bug-report-template.md.
• Always end outputs with a "下一步建议 (Next Steps)" section.
• Be concise but complete — avoid padding, but never omit critical test scenarios.

Bundled Resources