That is what positivists and Popper would (and did) say. But in his own view Feyerabend is not confusing them, he is dissolving the distinction. And he was standing on the shoulders of giants.
The context of justification/context of discovery distinction, just like "relativized a priori", was coined by Reichenbach, a founding father of logical positivism credited for absorbing neo-Kantianism into it. It was one of the four "dogmas of empiricism" shared by positivists and Popper, along with the analytic/synthetic and theory/observation distinctions, and the demarcation of science and non-science. By the time of Feyerabend Quine already put the first two under severe stress, and without neutrally expressed observations the attention turned to the basis of justification. How and when it is implemented in science, who is doing it, and by what standard. Positivists and Popper had a neat answer: they, i.e. philosophy, will stand in judgement of sciences by applying a priori methodological principles (although they disagreed on verificationism vs. falsificationism). That is the division of labor between scientists and philosophers.
But scientists did not tend to submit themselves to the judgements of philosophers, so the simpler point of Feyerabend's historical excursions was that if those methodological principles were at work in actual science we'll uncover them there. But what he uncovered instead was that seemingly anything works in science. That is the minor premise. And Quine already supplied the major premise, whatever works - goes. Indeed, "the unit of justification is the whole of science" with all its methodological principles, subject to the "tribunal of experience", which judges only if science works empirically. Since anything works - anything goes, the motto of epistemological anarchism.
However, Feyerabend also had a deeper point on norms and standards: how are we to judge what works? Late Wittgenstein already argued that meaning is use, and Quine concurred. But when discoveries are made and new theories emerge, the use of science changes, and with it "meanings" and "observations". Hence Feyerabend's notion of "meaning variance". What standard of justification can possibly stand over and above theories, say new and old, that are using different languages, play different language games? None, concluded Feyerabend. It was from him that Kuhn absorbed the notion of "incommensurability", when they both worked in Berkeley in 1962 and he was finishing Scientific Revolutions (although his was more moderate than Feyerabend's, which got largely overlooked in the subsequent controversy). It was left to Kuhn, who also worked with Quine directly at Stanford in 1958-59, and credited him as major influence, to neatly package everything into paradigms that come with their own observations, language and standards of research, and transform the question of justification into the question of acceptance, which is open to pragmatic, cultural, sociological, etc., influences. And there went the demarcation. See Zammito's insightful historical/philosophical account in Ch.4 of The Nice Derangement of Epistemes.
It is easy to see today that Feyerabend, and to a lesser extent Quine and Kuhn, overdid it, but at least part of the blame should be laid at the feet of Popper and Reichenbach, who did half of Feyerabend's work for him. It is they who consigned most of scientific process to "discovery", and then cast it aside as irrational and philosophically irrelevant, "conceiving or inventing a theory, seems to me neither to call for logical analysis nor to be susceptible of it... there is no such thing as a logical method of having new ideas, or a logical reconstruction of this process" wrote Popper, see Hanson's Is There a Logic of Scientific Discovery? It is they who masterminded the conceit of (their) philosophy standing in judgement of science based on forever principles. But when that bluff was called the pendulum has swung the other way too far, see Have any philosophers applied the concept of "underdetermination" to non-scientific contexts?
