菜单
Test AI-featured web apps with Playwright Python, including streaming, waiting, variable questions, resume, completion, and recovery.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | agent-webapp-testing |
| description | Test AI-featured web apps with Playwright Python, including streaming, waiting, variable questions, resume, completion, and recovery. |
Reuse Anthropic webapp-testing for browser interaction, logs, screenshots, and assertions.
Test Agent UI as state machine: running -> streaming -> waiting_user -> running -> completed.
Use templates/test_agent_flow.py. Fail after 15 seconds without visible text or state change.
Never locate Agent questions or answers by display text. Use stable data-testid and data-answer-id; text may vary by prompt, language, or model.
Save trace, screenshots, final response, and HTML QA evidence.
