Using something outside normal design parameters is an 'unbounded risk', that is there are too many unknowns to allow a prediction to be made. The absence of data makes repurposing the cylinder as a boiler component a bit dodgy.
On the plus side the cylinders are professionally made of steel and pressure rated at 20C to 100 bar (about 1400psi). That's substantially stronger than a typical small copper boiler.
On the negative side, we don't know how good the steel is at resisting hot wet corrosion, or if the construction is in any way vulnerable to heat. Also, the pressure rating is for a disposable item, ie. a gas cylinder not designed to withstand more than one cold pressure cycle in it's entire working life. As such the build is likely less strong than would be used for a hot boiler cycled several hundred times.
Although uninsurable my feeling is that the pluses outweigh the negatives as a way of occasionally making low pressure steam in a private workshop. Up to, say, 60psi with suitable precautions like a shield. As the cylinders are cheap the risk could be reduced by changing the boiler shell before fatigue or corrosion have had time to cause damage.
It would be interesting to comprehensively test a boiler made from a gas cylinder over a few months and then take it apart. If all was well inside, confidence would rise. Or the opposite!
Dave
Edited By SillyOldDuffer on 10/03/2019 17:32:13