> That's more specific than sneaking in with mRNA to get the body to produce something.
No, it really isn't more than that. You could send mRNA code into the body that produces a designer antibody that is specific to the cancer a patient has (similar to CAR-T, Keytruda, etc.), or you could design mRNA that specifically causes cancer cells to express antigens that attract the immune system. I'm obviously oversimplifying a pinnacle of human achievement, but on paper it should work that way.