AI-Powered UI Automation of E‑Commerce Checkout with TestHub Agent Browser Skill

Importing the Agent Browser Skill

The Hermes digital‑person module and WebSocket connection are assumed to be configured. The Agent Browser Skill zip file is downloaded from the TestHub VIP Skills repository and imported into the TestHub platform.

Case 1 – TestHub Open‑Source Version‑Number Creation

Task description sent to the agent‑browser skill (headed mode):

使用agent-browser skill，帮我完成以下内容的自动化测试（有头模式）： 
1.访问 http://localhost:3000，使用账号testadmin、密码admin123456进行登录； 
2.打开“AI用例生成”页面，切换到新打开的标签页，进入版本管理，点击“新建版本”按钮 
3.在新建版本弹框中填写以下信息： 
● 版本名称输入“V8.0”+当前时间戳； 
● 点开关联项目下拉框，选择下拉选项中的第一条， 
● 版本描述填写”这是 Browser Use 填写的版本号描述“ 
4.点击“保存”按钮，提交创建； 
5.验证版本管理列表第一条是否为新增的版本号。

Key observations:

Explicitly specify the Skill and request headed mode when required.

Use a strong multimodal model (e.g., Kimi‑2.6) for reliable execution.

The test executed successfully on the first attempt. The UI displayed a new version entry “V8.01786110889”, confirming the workflow passed.

Case 2 – E‑Commerce Platform Checkout Flow

Manual baseline steps (for reference):

Open the site and log in with credentials.

Navigate to the homepage and locate the computer category.

Select the product “MacBook Air 13 M3 16G (new)”.

Choose specifications (16 GB + 512 GB, silver) and click “Buy Now”.

Submit the order on the confirmation page.

Choose a payment method and complete payment.

AI prompt sent to the agent‑browser skill (headed mode):

调用agent-browser skill，使用有头模式，帮我完成以下内容的自动化测试：</code><code>1.访问 http://localhost:5175/</code><code>2.使用账号15252162668、密码123456登录；</code><code>3.点击电脑类目</code><code>4.点击查看“MacBook Air 13 M3 16G (新款)”详情</code><code>5.选择16G+512G、银色，立即购买</code><code>6.提交订单并支付

The TestHub platform executed the steps rapidly. Screenshots captured during execution show a successful checkout, confirming that the order was created and payment completed.

Key Technical Takeaways

Always highlight special commands such as the chosen Skill and headed mode.

Write task descriptions that mirror human actions step‑by‑step.

Prefer a capable multimodal model (e.g., Kimi‑2.6) for consistent execution.

The TestHub + Agent Browser Skill combination can be reused for test case generation, test plan creation, mobile‑app testing, and automated bug reporting.