A new shell is created for each command run by an AI model's Bash tool. When the shell closes, any programs started by that shell also close. This can make it seem like a program started but then crashed.
By Pawan Nayar · from the LLOS AI Learning record
A new shell is created for each command run by an AI model's Bash tool.
When the shell closes, any programs started by that shell also close. This can make it seem like a program started but then crashed.
An AI model's Bash tool starts a new shell for every command. A program started with `&` or `start /B` is linked to that shell. When the shell finishes its task, it closes. The program started by the shell then receives a signal to stop. This causes the program to close shortly after it appears to begin.
A developer asks an AI model to start a web server using a Bash command. The AI model runs the command in a new shell. The web server starts, but the shell immediately closes. The web server then stops, making it look like it failed to launch.
Resolved from this lesson's own record — real definitions, not invented for this page.
More from the same record, pulled by shared topic — not written new for this page.
When an AI model starts a program, the new program is linked to the AI model's process. If the AI model's process closes, the program it started will also close.
An AI model runs commands in a temporary shell. When the AI model finishes its task, this shell closes.
When a developer runs a Python script in the background using a shell command, the script is linked to that shell. If the shell session closes or times out, the operating system stops the script.
Before starting a server, a developer must kill all existing instances. After killing, the developer must check that no instances are running.
A new shell is created for each command run by an AI model's Bash tool. When the shell closes, any programs started by that shell also close. This can make it seem like a program started but then crashed.