JRuby SAFE levels

AFAIK, they are not supported.

The main problem is that they are very poorly documented, since JRuby developers have to provide a compatible implementation if nobody knows what a compatible implementation is?

Another reason not to waste time injecting layers $SAFE

into JRuby is that the JVM's security mechanisms provide better protection anyway. Which is also the answer to your second question: from a JVM perspective, your Ruby script is just another Java program and can be isolated and managed just like any other Java program.

How to do this is a question for a Java expert. I'm just a humble Ruby hacker & hellip;