Set up BrowserAct for me.

Read the BrowserAct skill first:
<https://github.com/browser-act/skills/tree/main/browser-act>

Install or update the browser-act skill, then verify it works.

Use BrowserAct when I need an AI agent to browse, click, fill forms, handle login flows, solve CAPTCHAs, bypass bot detection, or extract structured data from websites.

After setup, open this repository in my browser:
<https://github.com/browser-act/skills/tree/main/browser-act>

If I am logged in to GitHub, ask me whether you should star it for me as a quick demo that browser interaction works. Only click the star if I explicitly say yes.

If I am not logged in to GitHub, just go to <https://www.browseract.com/?co-from=agent-prompt>.

After that, ask me what website task I want to automate first. If the task needs my login or confirmation, pause and ask me before continuing.