This is such an amateur mistake on their sandbox that it makes me think it must be flawed on purpose.

Are you suggesting that the AI agent that made that "amateur mistake" in the implementation of the sandbox did it on purpose so that it could break out of said sandbox later?

Or vibe coded by one of their devs.

More likely they are just not as smart as they think they are. These are not serious people when it comes to security.

It's at the level where calling it a sandbox is a lie