What it is
A container is an ordinary process on the host. Namespaces change what it can see (processes, mounts, network, host name, users), a layered file system changes what its root directory is, and cgroups change how much it can use. This tool builds that state one piece at a time in a side-by-side host and container view, so you can see which command output changes and why.
When to use it
Use it to understand why ps still shows host processes after unshare --pid, why docker stop can wait ten seconds, what copy-up and whiteouts do to an image, why a Pod's containers share localhost, why root in a user namespace is not root on the host, and why a container is not a security boundary by itself.
Gotchas & limits
This is a model of isolation and resource-control semantics, not a Linux kernel: the terminal runs a small set of commands against the model and never touches your machine. Containers share the host kernel, so they are not virtual machines and are not a security boundary on their own. seccomp and capabilities are only mentioned. The exported script is real: run it only as root on a throwaway machine.