feat: Allow background command execution via optional parameter

[fede-ciliberti]

Is your feature request related to a problem? Please describe.
Currently, the execute_command tool is interactive by default. It waits for the first line of output and then prompts the user for action. This blocks the agent from performing other tasks while long-running commands (like a development server) are active.

Describe the solution you'd like
I propose adding an optional background boolean parameter to the execute_command tool.

When background is set to true:

• The command will run without requiring user interaction.
• The agent will be immediately unblocked to continue with other tasks.
• The command's output will still be streamed to the terminal UI.

This provides a simple and direct way for the agent to manage long-running processes without disrupting its workflow.

Example Usage

<execute_command>
  <command>npm run dev</command>
  <background>true</background>
</execute_command>

Describe alternatives you've considered
We considered collaborating on PR #6587, but our proposed solution is simpler and more direct for the agent to control. It adds a clear, explicit parameter for the desired behavior.

[roomote]

I'm exploring a potential fix for adding the background parameter to the execute_command tool. I'll work on implementing this feature and open a PR to share progress.

[roomote]

Opened PR #9023 with my attempt at addressing this issue: #9023

The implementation adds the optional background parameter as requested, allowing commands to run without blocking the agent while still streaming output to the terminal UI. All tests are passing and the feature is ready for review.

Feedback welcome!