user-simulation-test

chkim-su 的头像作者 chkim-su

端到端测试中模拟真实用户行为的规则和模式。涵盖UI交互、API流程和基于场景的测试策略。

📖 中文导读:端到端测试中模拟真实用户行为的规则和模式。涵盖UI交互、API流程和基于场景的测试策略。 以下正文为该技能的原始 SKILL.md 内容(英文),上方安装命令可直接使用。

description: Rules and patterns for simulating real user behavior in E2E tests. Covers UI interactions, API flows, and scenario-based testing strategies. name: user-simulation-test allowed-tools: ["Read", "Grep", "Glob", "Bash", "Task", "mcp__plugin_playwright_playwright__*"]

User Simulation Test Rules

Core Principle

Test as users do, not as developers think.

Simulate complete user journeys including:

  • UI interactions (clicks, inputs, drag-drop, scrolls)
  • Form submissions with realistic data
  • Multi-step workflows with wait states
  • Error scenarios and edge cases

Test Type Selection

Project Type Primary Test Secondary Test Tools
Web Frontend Playwright E2E API Mock browser_* MCP tools
API Backend HTTP Sequence Unit Tests curl/httpie + project runner
Full Stack Integration E2E API + UI Playwright + API calls
CLI Tool Script Execution Unit Tests Bash + project runner

Scenario Structure

Every test scenario must include:

Scenario:
  name: string           # Descriptive name
  description: string    # What this tests
  preconditions: []      # Required state before test
  steps: Step[]          # User actions in sequence
  assertions: []         # Expected outcomes
  cleanup: []            # Post-test cleanup actions

Detailed templates: Read("references/scenario-templates.md")


UI Interaction Patterns

Click & Input

Action Playwright Tool Notes
Click button browser_click Use ref from snapshot
Type text browser_type Set slowly=true for key handlers
Fill form browser_fill_form Batch field updates
Select option browser_select_option For dropdowns
Press key browser_press_key Enter, Escape, etc.

Advanced Interactions

Action Playwright Tool Use Case
Drag & Drop browser_drag File upload, reordering
Hover browser_hover Tooltips, menus
Wait for text browser_wait_for Loading states
Take screenshot browser_take_screenshot Visual verification

API Testing Patterns

HTTP Sequence Test

APISequence:
  - name: "Login"
    method: POST
    endpoint: /api/auth/login
    body: { email, password }
    extract: { token: "response.data.token" }

  - name: "Create Resource"
    method: POST
    endpoint: /api/resources
    headers: { Authorization: "Bearer ${token}" }
    body: { ... }
    assertions:
      - status: 201
      - response.data.id: exists

Test Data Strategy

Realistic Data Generation

Field Type Strategy Example
Email Pattern + timestamp test_${timestamp}@example.com
Name Fake names Korean/English name generators
Phone Pattern 010-XXXX-XXXX
Address Realistic fake Valid format, fake data
Payment Test cards Stripe/Toss test numbers

Data Isolation

  • Always use test-specific data prefixes
  • Clean up created data after tests
  • Use separate test database/environment when possible

Wait Strategies

Situation Strategy Tool
Page load Wait for element browser_wait_for text
API response Wait for network idle browser_network_requests
Animation Fixed delay browser_wait_for time
Loading spinner Wait for disappear browser_wait_for textGone

Assertion Patterns

UI Assertions

UIAssertions:
  - type: text_visible
    value: "Success message"

  - type: element_exists
    selector: "[data-testid='result']"

  - type: url_contains
    value: "/dashboard"

  - type: no_console_errors
    level: error

API Assertions

APIAssertions:
  - type: status_code
    expected: 200

  - type: response_time
    max_ms: 2000

  - type: json_path
    path: "$.data.items"
    condition: "length > 0"

Error Handling

Graceful Degradation

  1. Screenshot on failure - Capture visual state
  2. Console log capture - Get browser errors
  3. Network request dump - See failed API calls
  4. State preservation - Keep test state for debugging

Recovery Actions

OnFailure:
  - capture_screenshot
  - capture_console_logs
  - capture_network_requests
  - preserve_browser_state  # Don't close immediately

Test Categories

Category Frequency Coverage
Smoke Every change Critical paths only
Regression Pre-release Full user journeys
Integration Feature complete Cross-service flows
Load Scheduled Performance under stress

Test patterns: Read("references/test-patterns.md")