Devin signs in once and remembers. Build a QA agent that tests behind your login with the Claude Agent SDK

说明与用途

使用已保存的浏览器配置文件并从保险库中安全检索凭据的 QA 代理，加速测试并保护密码不被模型上下文泄露。

问题

QA agents that log in on every run waste time on login pages and expose passwords in prompts.

结果

The QA agent runs quickly, never re‑enters passwords, and maintains authenticated state across CI runs.

工作流

• Create a Burrowbox machine with "browser":"full" to preserve sign‑in state across restarts.
• Store the staging test account credentials in the machine’s vault using the browser_login tool.
• Write a QA script that starts the machine, waits for it to be running, then uses the Claude Agent SDK to navigate the checklist, calling browser_login only when necessary.
• Run the script in CI, letting the machine keep its state for 30 minutes before stopping automatically.

证据

A person logs in once through Devin's interactive browser, in the session's Desktop tab, and handles SSO redirects, MFA prompts and CAPTCHAs themselves. When they ask Devin to "save the browser profile", Devin zips the browser's cookies, localStorage and other Chrome profile data and attaches it to the organization's blueprint.

来源

The QA agent runs in your CI job with the Claude Agent SDK. It connects to a Burrowbox machine's MCP endpoint, works through a checklist in that machine's browser, and finishes with PASS or FAIL.

来源