Interesting. I see it the opposite way. People are proposing all kinds of one-off hacks for their specific use cases, which in my experience leads to more bugs and less usability, and exponential complexity as they don't work together. I would rather see "what if we had full scripting again?" and review proposals from that angle.

As to safety, that still needs to be assessed. Details really matter, so I cannot give a verdict until the design and implementation are complete. So step 1 is to do that. Step 2 is to evaluate whether it's worth it.