Daniel W. Drezner is a professor of international politics at the Fletcher School of Law and Diplomacy at Tufts University and a regular contributor to PostEverything.

National security adviser H.R. McMaster sits behind President Trump during a Cabinet meeting at the White House in December. (AFP)

A year ago, the name H.R. McMaster would have evoked near-unanimous praise from the national security community. He was a venerated tank commander and counterinsurgency warrior who earned a PhD in history. He wrote a widely praised book about the failure of military leaders to challenge civilian leadership on Vietnam War strategy. And although no one was thinking about it in mid-January, he seemed like a much better person to advise a president than, say, Michael Flynn.

That was then, and now McMaster is Trump’s national security adviser. This would have been a challenging job for anyone given the current president’s … um … let’s say “decision-making style.” Last month McMaster went on the record to acknowledge that on foreign policy, Trump “has moved a lot of us out of our comfort zone, me included.” Reports about friction between McMaster and Trump have fueled widespread speculation about his departure in 2018.

Perhaps, then, it is a good idea to consider how well McMaster has been doing at his current job. Based on recent news reports, there’s some good, some bad and some ugly as well. In order:

The good: According to the Daily Beast’s Betsy Woodruff and Spencer Ackerman, McMaster is working hard to make sure Trump does not scuttle the Iranian nuclear deal. At the same time, he is trying to do right by his boss:

According to multiple sources, H.R. McMaster is reprising the role he played last fall: removing a legislative irritant from Trump so that the president can quietly remain in the deal.

This time around, sources told The Daily Beast, McMaster is searching for an agreement, even one just in principle, with the leadership of the Senate Foreign Relations Committee to convince Trump that the president’s decision to “decertify” Iran’s compliance last fall pressured Congress to modify its terms.

It doesn’t exactly work like that — Congress is not a party to the Iran deal, and so all it can do here is place restrictions, or encouragements, on U.S. policy toward Iran. But multiple sources said McMaster considers that an agreement with Sens. Bob Corker and Benjamin L. Cardin that took away the congressional deadlines on Iran that Trump hates would be a face-saving way for the White House to accept the deal.

This is an example of good staffwork. McMaster is trying to please his boss without making a volatile Middle East even more unstable. Beyond Trump, no other significant player is interested in nixing the deal. None of the other signatories want the nuclear deal (JCPOA) to be abrogated. Iran has signaled that cooperation with the IAEA might be scuttled if Trump ends U.S. participation in the JCPOA. Staying in the deal makes little foreign policy sense; the Iran protests do not change that fundamental equation.

Getting Trump out of having to waive sanctions might seem like a large expenditure of time and effort for a picayune reward. Still, this is a case where the national security adviser is trying to thread a needle between good foreign policy and the whims of his commander in chief. If McMaster cannot convince Trump that waivers and certifications are no big deal, this is the next-best thing he can do.

The bad: McMaster’s ultra-hawkishness on North Korea continues to make little sense. The Atlantic’s Uri Friedman took a long look at McMaster’s views on deterrence that is well worth reading. This part stood out:

As McMaster tells it, traditional deterrence (if you strike us, we’ll strike you), which helped dissuade the Soviet Union from firing nuclear weapons at the United States during the Cold War, may not work with a government as brutal as Kim Jong Un’s. If the world’s most despicable regime isn’t prevented from acquiring the world’s most destructive arms, what’s keeping other nations from racing to build their own nuclear arsenals? North Korea, which has exported missiles and nuclear-related materials to countries such as Iran and Syria, could sell nuclear weapons to America’s enemies, he warns. And a nuclear North Korea could blackmail U.S. leaders by, for example, threatening to incinerate Los Angeles unless America withdraws support for its ally South Korea, exposing the South to invasion by the North.

In brief, North Korea’s development of a long-range nuclear capability “would be the most destabilizing development … in the post-World War II period,” McMaster says.

The only part of that excerpt that makes sense to me is McMaster’s concern about proliferation. North Korea as a proliferating actor is a terrifying thought, and may well require military action. U.S. intelligence failures in this area are also disturbing.

That said, McMaster’s belief that deterrence will not work on Kim Jong Un is incomprehensible. There is little evidence of Kim’s irrationality. He wants different things than most American foreign policy advisers; that is not a sign of mental instability. It is worth remembering that dictators with far more power and far less mental stability — Stalin and Mao — controlled nuclear arsenals far more vast than Kim. Deterrence worked in those cases. And it is worth noting that the alternative scenario of a “bloody nose strikemakes zero sense.

So does the idea that the DPRK’s success in developing nuclear weapons will encourage other potential aspirants to do the same. North Korea paid a high price in terms of lost economic output to develop its nuclear and ballistic missile capabilities; it succeeded because it was willing to forgo the benefits of exchange with the outside world. Etel Solingen suggests that states interested in engaging the global economy can be pressured into abandoning nuclear programs. It is difficult to think of countries beyond North Korea that are willing to do that.

I’ve said it before and I’ll say it again: I do not understand McMaster’s thinking on this issue.

The ugly. The Washington Post ran a long story last month about Trump’s failure to check Russian threats to the United States. It included a disturbing anecdote involving Trump, McMaster and Fiona Hill, the NSC director for Russia:

In one of her first encounters with the president, an Oval Office meeting in preparation for a call with Putin on Syria, Trump appeared to mistake Hill for a member of the clerical staff, handing her a memo he had marked up and instructing her to rewrite it.

When Hill responded with a perplexed look, Trump became irritated with what he interpreted as insubordination, according to officials who witnessed the exchange. As she walked away in confusion, Trump exploded and motioned for McMaster to intervene.

McMaster followed Hill out the door and scolded her, officials said. Later, he and a few close staffers met to explore ways to repair Hill’s damaged relationship with the president.

Every time I read this anecdote, I try to figure out why McMaster initially scolded Hill rather than explain who she was to the president. Perhaps McMaster surmised that once Trump perceived insubordination, persuasion would have been difficult. Still, acting like Hill made the mistake in this situation was poor leadership on McMaster’s part. Pleasing superiors is important, but so is protecting underlings. He appears to have failed in the latter task.

I can only imagine the pressure McMaster has faced with a mercurial president who needs lots of “Executive Time” to manage. Still, some of what the national security adviser is thinking and doing remain inexplicable.