BrowserSkill Made Browser Automation Feel Surprisingly Normal

Description and use

BrowserSkill lets AI agents automate tasks in the user’s existing browser session, offering visibility, control, and security.

Problem

Traditional browser automation requires separate profiles, cookie copying, or fragile auth flows, disrupting daily workflow.

Outcome

Agents can automate repetitive web tasks—dashboard checks, data collection, UI issue reproduction, form filling—within the user’s trusted browser session, improving productivity while maintaining control and security.

Workflow

• Clone the BrowserSkill repo: git clone https://github.com/Tencent/BrowserSkill.git
• Navigate into the repo: cd BrowserSkill
• Follow setup instructions to install dependencies and build the extension
• Connect the extension to the local CLI
• Configure browserAutomation settings in the agent’s config to enable automation and set confirmation gates
• Run the AI agent; it will perform actions in the existing browser session while allowing user inspection and control

Evidence

BrowserSkill’s CLI-and-extension approach feels much closer to a practical daily tool for Cursor and other shell-capable agents.

Source

Add confirmation gates for destructive actions, uploads, purchases, and messages.

Source