That's not so long ago. (Since late 1950s).
You should not compress the whole nuka-ball. You should just slap its shell to make it slap the less-than-kg sparkplug (and sometimes the inert betatronesque sphere around it to force the Xray emission), it gets slightly supercritical and ignites the puny amount of gaseous D+T inside, and then the staged process of fusion-fission-fusion-fission runs even without critical mass of the charge in whole.
You need just several kg of the chemexplosives for that (thanks to the Pu's delta-alpha crystal phase transition).
Also, usage of U/Pu deuteride helps since the Ray&Ruby nuka-tests.
If you have enough purified Pu, you can make a briefcase-sized linear (gun-type) charge of at least two schemes.
But it's very expensive, as you have to take the Pu very early from the reactor, before it's filled wth super-active parasite isotopes, and then expose it for a decade to let them decay.
But there is no problem about ten-kg nukes, except their cost and applicability in whole.
The main safety system of them is their tritium addiction. Hard to create, easy to burn out.
Actually, you couldn't make a simple nuke out of a normally designed nuke. It doesn't contain the critical mass of isotopes.
Wait... Oh, ...!
https://wiki.kerbalspaceprogram.com/wiki/Kerbin
Density
58 484.090 kg/m3