From the course: AI Coding Agents with GitHub Copilot and Cursor

Unlock this course with a free trial

Join today to access over 24,700 courses taught by industry experts.

Running agent command line prompts

Running agent command line prompts

From the course: AI Coding Agents with GitHub Copilot and Cursor

Start my 1-month free trial Buy for my team

Running agent command line prompts

“

- [Instructor] In addition to generating, modifying and deleting code, files and folders in your project, the AI coding agents in GitHub Copilot and Cursor can also run commands in your terminal. This is where the human-in-the-loop component becomes essential, because once the agent can access your terminal, it can run commands, access external services, change settings on your computer, boot up servers, download content and a bunch of other things. And all these things are really useful, but you have to know what is happening and therefore both GitHub Copilot and Cursor's agent require you to click buttons and say, yes, I want to run these commands, when this happens. Based on my experience, where you'll encounter this the most is when the agent determines you need to install some sort of dependency, because this is all done through the command line. So let's see if I can trigger this behavior in GitHub Copilot to give you a demo of this experience. In my project, I'm using Shadcn…

Contents