I've always asked the product manager (hopefully a single person, not a committee), whom I understand to be the person who decides which qualities (like performance, usability, security, and what @ripu1581 mentions) are important to the product and who decides which features are developed. If that responsibility lies with somebody titled "Architect" or "Developer", that's fine as well. You need to synch with this person to ensure results from the test are implemented and thus, will become visible.
If you have knowledge in the business area and with previous version of the product, they are usually happy to consider proposals for tasks from you. If you are new to the team, you probably cannot press your own agenda of what should be changed, but need to follow their ideas of what is important. (That's not bad, since it's usually taken me a single usability test to surprise the team what users find intuitive and what not - so after the test results are out, your opinion will be appreciated much more.)
I would not extend the scope of the test across what is feasible to tackle in the near future. Preparing and conducting the test is effort for everybody, and the closer the connection between this effort and immediate changes, the clearer the value of a usability test becomes.
In addition to the tasks, you most likely also need a screener: A document describing your target group, which allows you to sort out people who your product is not addressing. Be careful when preparing this with the product manager! If anybody wants to question your results, a marginal fit between target group and test participants is their best argument.
To ensure everyone is informend, invite everyone from developer to manager to tester to watch the sessions - stories about users being stuck stay with the team and help you in subsequent priority settings.