But it's not tit-for-tat from the U.S. side yet. The proposal is that the U.S. is not engaging significantly, and that a western country nukes 1+ targets in order to bring it into the fight. We are sitting on our hands, then suddenly a thermonuclear explosion happens against U.S. interest. That's going to garner some interest from the interagency.

Also, the instruments of national power that would investigating are, in Twilight 2000 terms, CIVGOV side. While DOD might be drafting, training and deploying troops to respond (or preparing to respond) conventionally to an Article 5 invocation, DOE, FBI, etc. would be working the forensic side of any nuclear attack against the U.S. forces/fleets/bases/cities. I'm not saying they would be bored and looking for work (CI issue would occupy a lot of the FBI's focus in a pre-war/wartime environment), but the Federal bureaucracy *can* walk and chew gum at the same time.

Now if multiple nuclear attacks are employed simultaneously or in rapid succession, that could make forensic analysis difficult/moot. However, that dramatically increases the possibility of launch detection, leaks, SIGINT interception, etc.

Can something be mistaken at the time of launch for something else? Sure. Yeah, some guy training never to land an aircraft, just to fly one, might not raise any red flags at the time. But investigations like this work backward from a known event. Logs will be reviewed, radar/sat log/data files examined, people interviewed. The goal is high confidence attribution. Yes that takes time, but again, there's no reason it won't happen simultaneously to any DOD activities.

PS: Cruise missiles would avoid a lot of the detection issues, but striking a U.S. base or city still invites technical analysis. That debris remains. Striking a fleet at sea however...