It is perhaps possible, but currently neither quick, reliable nor cheap. You'd need to check that: (a) it's not immediately toxic, (b) there's no immediate metabolisation that renders it inactive, (c) it can enter the bloodstream, (d) it can penetrate the blood-brain barrier, and (e) it can either substitute for a particular neurotransmitter, act as a releasing agent, act as a reuptake inhibitor, or several other proposed and known methods of actually causing 'mind-altering' effects.
This is essentially the 'drug-discovery pipeline' that big pharma spends hundreds of millions on when developing new medications.
I think you're taking it a bit too far. If say i sell bath salts, which aren't supposed to be consumed, why would there be a need to check metabolism/toxicity and such ?
But if metabolism can transform said chemical , then sure, you've got to take it into account. But many of those drugs can be detected by pre-metabolism binding, i think.
Metabolism can go either way, for instance Codeine is thought to be mostly a prodrug[1], which is converted to morphine in the liver. So it's only really effective when taken orally, because it relies on the first-pass metabolism for its action. Morphine itself, on the other hand, is quite significantly converted into inactive metabolites when taken orally. So you end up with IV morphine being ~3x more active as the same oral dose, because it bypasses the first bit. Other drugs are exclusively one route because they rely on other processes for transformation or absorption.
This matters because it's not really appropriate to schedule chemicals which happen to trigger a response in a test-kit, but are not plausibly drugs of abuse due to side-effects, or lack of any reasonable mechanism of getting to the sites where they'd be active.
The "intended for consumption" is a convenient fiction for the head-shops and dodgy websites that sell such things, but actually having to test things which aren't intended for use in certain ways would be a massive imposition on, well, everything that uses chemicals[2].
There are heuristics that cover a lot of ground, and that's basically what the Analog Act tries to preempt, but there are still huge numbers of potentially active compounds we have no idea about. In-vitro binding studies are perhaps the next filtering step, but you still ought to have to demonstrate actual demonstrable plausibility of action for it to be a crime[3]. In addition, if you have something that feels like cocaine, but rapidly metabolises into a lethal neurotoxin or whatever in the same dose, it's not going to be a drug of abuse, except perhaps in suicides.
In terms of banning/restricting sale of things, maybe binding data would be sufficient, but I'm not sure just what the implications for testing & compliance for everyone involved would be.
[2] Despite some TV adverts to the contrary, this would be everything.
[3] Leading to (IMO) the utterly bizarre world of fakes/illegal placebos, for which in many places you can be convicted for the same degree crime as the drug you purported to be selling, rather than say, fraud.
This is essentially the 'drug-discovery pipeline' that big pharma spends hundreds of millions on when developing new medications.