The quote is direct from Intel. Shorting power supply outputs must
not damage power supplies.

Either power supply provides power just fine - without damage - or
power supply enters foldback current limiting. Standards even provide
volt-current graphs for all currents - for all loads. If a power
supply provides overcurrent, then it does so without damage or it
protects itself - enters foldback current limiting. But again, nothing
new. This required operation has been standard for 30+ years. Where
overcurrent causes overheating and power supply damage, start with the
most likely reason for failure: human who is a 'bean counter'; not
technically knowledgeable. Market is ripe with power supplies missing
essential features because so many computer assemblers do not even know
about foldback current limiting. They buy on dollars; don't ask for
spec sheets. Power supplies dumped into a market of computer
assemblers who don't even know how electricity works.

If a power supply provides too much current, then it does so without
damage or it enters foldback current limiting. Furthermore, circuits
don't care where that current is going - to ground or where ever.
Current is measured. If current is too high, then foldback current
limiting takes over. But then this is simple fundamental stuff made
obvious even with simple power supply design knowledge - stuff I was
doing even as a teenager.

A short from any one voltage to any other voltage must not cause
damage: is required by industry standards, as is required by Intel
specs, and as was defacto standard 30 years ago.

A quote is direct from Intel:
> The power supply shall be capable of withstanding a
> continuous short-circuit to the output without any
> damage or overstress to the unit ...
Could not be any clearer and blunter.

BTW, when I designed power supplies, that requirement was easily met.
Those requirements are basic stuff. If one cannot understand defacto
power supply standards of 30 years ago, then how can that person answer
the OP's question? Virus does not damage properly designed hardware.

Robert Mabee wrote:
> Foldback current limiting is safe into a dead short to ground, but
> not necessarily into a continuous overcurrent at nearly normal voltage.
> Damage in that case is thermal (slow) or results from a transistor
> switching off so slowly that an effective short exists across a
> capacitor (instant death for a chain of components), or happens in the
> wiring rather than in the components.
>
> A short to a different voltage output is quite likely to fry something
> as most outputs do not have overvoltage protection or diodes to prevent
> reverse voltage. Even perfect protection at the supply would let the
> wires of a low current output melt if shorted to a high current output.
>
> Any short applied after voltage is established will cause widespread
> damage as the total energy stored in capacitors is transferred to the
> wire path.