"What should I do?" "What should one do?" "What would a virtuous person do?"
There are many ways to map the terrain of ethical theories. One that I have found useful follows the kinds of questions we commonly ask when considering moral questions.
"What should I do?" is a practical but not yet an ethical question because one can give a nakedly egoistic answer: "I will steal a car to get to my destination; I will do whatever it takes to get rich or to save the planet."
"What should one do?" is an ethical question, because (paraphrasing Hume) by abstracting from "I" to "one" the speaker departs from his particular situation and chooses a point of view common to him with others. The relevant commonality is a feature which everyone in the intended audience shares, a dimension of equality within which we can stand in the place of another. Trading places within a dimension of equality is the core operation of universalization, the main engine of normativity underlying most major modern ethical theories. Where those theories differ is in the dimension of commonality they consider most relevant.
The relevant dimension of commonality can be a feature of the real world: There are certain things that we all want--life, food, shelter, happiness.
Alternatively, the dimension of commonality can be a feature of our consciousness, specifically, our reasoning: we all share perceptual, logical, grammatical and linguistic concepts.
Each approach has its characteristic strengths and weaknesses.
The "real world" approach taps into a deep well of experiential plausibility (most would agree that pleasure is better than pain) but faces measurement issues (does one unit of pleasure equal one unit of pain?); aggregation problems (does my pleasure offset your pain?); discount rate concerns (by how much does future pain offset present pleasure?); border disputes (whose interests count? all humans? all sentient beings?); and questions about differentiated care (do everyone's interests count equally or should those close to us count more)?
The "reasoning" approach can dispense with much of the factual messiness and focus on the process and the logic of moral reasoning. Right is what any rational being must necessarily assent to. Rationalists thus derive moral truths from quasi-geometrical proofs (equal and non-rivalrous spheres of negative freedom can be scaled indefinitely), agreement (what everybody can agree to must respect everyone's freedom), recognition (I need you as an equal to fully be me), and performative consistency (using words such as "should" already commits the speaker to the principle of universalization). The rationalists' main problem is that abstract features of reason or discourse don't generate much motivational pull.
There is another, older path leading from the pre-moral "What should I do?" into the realm of ethics, namely the question: "What would a virtuous person do?" Instead of the abstract "one", featureless save for a dimension of commonality, the relevant benchmark is now a semi-real person, someone embedded in relevant real-world circumstances who's wise, just, benevolent, courageous and capable of moderation. Commonly it is even someone with a name: Hercules, Socrates, Kongzi, Mark Watney, etc. Instead of providing reasons for actions applicable to any "one" in the same situation, the focus is now on what kind of person specifically I want to be. That opens up different wells of normativity: the roles we play (father, child, spouse, colleague, citizen); the character traits we admire (honesty, courage, wisdom, kindness, moderation); what makes for a meaningful life (service, self-realization, altruism). This is the realm of virtue ethics. One of its unique strengths relative to its modern cousins is motivation. To want to be a certain kind of person is to want to act like that person. Virtue ethics also seeks to transform the actor into someone for whom acting wisely, justly, courageously, and with moderation in any given situation has become second nature, which is a more ambitious program than providing reasons for action to anyone, regardless of their character. The main weakness of virtue ethics, greater specificity and more limited reliance on universalization, is also the source of its strength.