Hidden commands were added to the Safe-Tcl security model. Instead of removing unsafe commands from an interpreter, they are hidden. The master can invoke hidden commands inside a slave. This is necessary so that the command sees the correct context. This adds new operations to the interp command: invokehidden, hide, expose, and hidden. Hidden commands are described on page 281.
Initialization of a safe interpreter with a safe base that supports auto loading and a standard exit alias has been abstracted into a Tcl interface. The safe::interpCreate and safe::interpInit procedures create or initialize a slave with the safe base. The safe::interpDelete procedure cleans up. The safe base is described on page 284.
To support the Trusted security policy, the interp marktrusted command was added. This promotes an unsafe interpreter back into a trusted one. Of course, only the master can do this.