No Bob, You Did Not Just Invent Bread

 Has something like this:

Player: So I'm casting Create Object to make 25 pounds of silver. This costs 4 energy, because I have skill 15, and gives us $25,000. We get $5000 each to spend.

GM: You know that it's not real money, right? As soon as they aren't touching it it vanishes?

Player: Yeah, but we just have to pay and leave before that happens. How would they ever know?

Ever happened to you?

I can help!

Exploitation Exasperation

Often players will come up with a use of magic or technology that seems really too easy. This is especially true when using an improvisational magic system like Ritual Path Magic or Sorcery, but can happen really with any capability the players have that doesn't exist in reality. This can create a problem of verisimilitude when the proposed use describes an exploit that wouldn't allow the setting to function as observed. In the example above, if merchants were unprepared for illusion magic, then they would already have been robbed by magical thieves, which doesn't describe a stable market environment that has the kinds of shops the players are about to try to rip off.

Exploits of this kind can be unexpected, and it is impossible to design a setting that isn't vulnerable to them in advance; not while still allowing the players to be creative with abilities. You can't have foreseen every possible creative use of spells, powers or technology, and have already designed something that addresses it. Improvisational worldbuilding is the answer here. Like many things with improvisation, this seems like it might be hard and require very quick thinking, but by keeping a few basic principles in mind it becomes a lot easier.

No Bob, You Did Not Just Invent Bread

1. The Old Hack

Unless the players truly have unique capabilities that have never existed before, you should assume that someone has already tried this. They aren't the first to do it, and it probably isn't worth counting how many people are ahead of them in trying it either.

2. Status Quo Inertia

Enduring institutions  generally want to keep things the way they are, because that's how they remain enduring institutions. The systems of the world are designed to be hard to disrupt and many people are deeply invested in keeping them that way.

3. Simple, Cheap and Easy Countermeasures
Whatever countermeasures that people use, they are going to be the easiest and cheapest ones to implement that still work. If it can be done with everyday tools and simple behaviors it will be.

4. The World Has Time And All The Money
While the players' resources for shenanigans are limited, the setting has effectively unlimited resources to prevent shenanigans, and civilization had a head start of thousands of years before the PCs were even born. Don't be afraid to retroactively say that expensive defenses built over centuries are in place if it makes sense that there would be.

5. Nobody Knows How Anything Works
You don't actually need to know all the details about how a countermeasure or obstacle works, the characters wouldn't know either. It's enough to know that they do work. Think about how credit card theft prevention, or public key encryption works in the real world. People are perfectly willing to trust complex systems they use every day without any idea how they were constructed.

6. The Characters Know Things Even The GM Doesn't
The characters exist in the world, and people in the world would know all kinds of details the GM hasn't thought about. You can tell the player that they already know that this exploit doesn't work. This is more fair to the player than just slapping them with consequences after the fact.

7. The Soulless Minions of Orthodoxy Get a Roll Too
Often the solution is just make it an opposed roll. This abstracts whatever systems exist to prevent the exploit into a quick and fair seeming mechanical solution. 

8. The Scale of Scam is Inversely Proportional to Scam Success
Ordinary people might not have many safeguards against someone using magic to scam them out of buying drinks, but banks certainly will against magic being used to scam them out of millions. Some tricks that work fine for small stakes won't work for larger ones.

9. Shenanigans Are Illegal Here
Typically a society's easiest response to things they don't like is simply to outlaw them. You don't need to have the entire legal code of a setting fully written in advance. If the thing the player wants to do sounds like it should be illegal, tell them that it is.

For the example above, you can point out that of course everybody knows you can use magic to counterfeit money so merchants always accept payment in the traditional brass bowl before touching it. You can tell the player that their character already knows this, so they wouldn't even attempt this scam in the first place. Especially since they also know the punishment if you are caught is having your Magery drained permanently.

Or if the players want to create a Ritual Path Magic spell that instantly kills the BBEG and saves the day without leaving their house, you can simply point out that they don't even know what magical defenses the villain has, but it doesn't work. You don't even need to figure out what those defenses are, it's more plausible that they exist and that's enough.

A ritual intended to disrupt the climate or some other global phenomena could run into needing to overcome the powerful wards that the guilds have constructed over centuries, which collectively resist world changing magics with effective skill-29.

But We Want To Change The World

Obviously sometimes people can do things that have never been done before. Everything happens for the first time, once. The important thing, for both verisimilitude and for satisfactory narratives is that this should generally never be easy. If you are going to use magic to rob the richest bank in the world, you should probably be able to try, but this isn't a single casual act. This is a heist and it requires research, planning, risk, and luck. It should be an adventure itself, not a single roll or spell while doing other things.