I'm going to go with no.
From experience, you normally get the most accurate and truthful user testing results by getting real users to do real tasks in relevant contexts. If you ask someone to want something they don't want, or behave in a way they don't normally behave, you get weird imagination interference where the user will try to emulate their idea of something rather than act based on the real desires.
It's the same reason why it can be hard to test with dummy content or in a lab. I think it's totally fine to ask someone to complete a task but in your example, I think the second example is much more effective as it doesn't dictate the context. Additionally, you're going to get better results when you show that stripped-back scenario to some users from different demographics because rather than declaring who they are, you can see how a range of different people go at achieving this task.
If you're asking that question with "you are an x" then you may find that you have only designed your app for x. If that's your intention, great, but remember that lots of different people use software, including users with accessibility requirements, users of all different ages and users who do and don't understand the basics of what your software does.
Don't box yourself in. Go broad, and you'll get real information from different people with different needs. That's the reality of how your software is going to be used, so don't be too specific and lose the important info that might come from widening your boundaries a bit.