So I have been theory crafting in the lead up to the release of Hemlock Vale, and no matter how I look at it, this is a very soft weakness. Like all weaknesses, timing is everything and this can sucker punch you at the end of the first upkeep. But Kohaku has so many tools available to deal with this, that its kinda silly. So without seeing everything in the box we already have:

Spectral Razor (0). An absolute Mystic standby. This wins against anything except the autofail on standard even in The Forgotten Age. If you already have curses in the bag, you still perform the test anyway, so the second worst case scenario is its 2 action, 1 card, 2 resources, and 2 horror. Pretty standard expectations for sig weaknesses Stats: ---2 actions (1 for the engage, because it is aloof. [I haven't been able to confirm if there is a ruling for myself that this lets you engage an aloof enemy. Rules as written you cannot, and the event is already powerful. But if you do play it as engaging an aloof enemy, cool. 1 less action.]) (Don't come at me. The wording on this card annoys me) ---1 card ---2 resources ---A base 7 to 2 test.

Manipulate Destiny (2). This needs experience so probably not your scenario 1 method. But if you're leaning hard into curse and have Favor of the Moon or your incredibly amazing signature in play, you're all set. This spell (Arcane Initiate anyone?) is not a test notably, and so ignores the specific wording on the Yurei that specifies during a skill test. For my money, A way to kill something that is aloof without engaging, and no test? Love. It. Stats: ---1 action ---1 card ---1 resource ---No engage required ---Testless

Drain Essence (0). A Hemlock card. A parley test at base 4 to 2, that does not require engagement, and can heal you. While you need a way to boost by 2 or 3 (hello Cat Mask/Holy Rosary/David Renfield), this is easily done. This is a test, so there is some risk, but the upshot is that you heal yourself by 2, and don't need to engage. It's all gravy. I'll take those odds. And depending on how much parley support we see in HV, running Fine Clothes, because you want to run Power Word on Kohaku and maybe another couple cards you will have access too, can turn this into an all but guaranteed success, with a potential price. Stats: ---1 action ---1 card ---2 resources ---Base 4-2 test, requires boosting or Fine Clothes ---No engage required ---It heals you assuming you have damage

Occult Lexicon/Blood-Rite (0). A classic standby for seekers normally, but because of past design choices, Kohaku happens to get one of the best cards in the game, albeit in a very competitive slot for him. When you don't need it, this does somewhat act as deck filtering, because you can always play the event and not discard cards, so you net draw cards, or you can discard duplicates to gain resources. Lots of flexibility, low downside, except that it's a one of in your deck, so reliably drawing it is awkward. minimal downside. Pretty comparable to Manipulate Destiny. I would personally prefer to upgrade out of Occult Lexicon, but your deck, your choice. Stats: ---2 action (1 to play the lexicon, 1 to play a Blood-Rite) ---1 card (technically) ---2 resources ---Testless ---No engage required

Gaze of Ourash (2). Overkill. Great if Weeping Yurei spawned at the same time as another enemy from the encounter deck. But since you'd rather be as efficient as possible and not rely on damage events, this feels like too big a chance to miss, even with your signature or favor of the moon. I feel like you could potentially have a higher ceiling with this, so it depends on the setup. If you were playing main fighter, I could see this having a place generally just as an AoE attack. But solo? Manipulate Destiny saves 1 resource and only uses 1 trigger of something to guarantee the hit. Stats: ---1 action ---1 card ---2 resources ---Testless ---No engage required

LATER EDIT: So I completely forgot about Nepthys and was previously focusing on the Cursed side. But payoff for going more heavily on the blessed side is similar to ||Zoey and Sister Mary, in that Nepthys if set up, can actionlessly and testlessly kill the Weeping Yurei. All for doing what you already want to do. So actually, if you're going more of a bless fighter, just take Nepthys, no brainer. Easy win; what weakness?

Honorable mentions.

Ofuda (0). Initial impressions are for me lukewarm on this because it feels like scenario tech. But for Kohaku who has an inbuilt reliable way of revealing blesses, this does remove elusive and aloof as keywords. This doesn't kill Weeping Yurei, but it does help your Armageddon kill it. But you could just take the engage action. This costs you 2 resources and a card. I'm not strictly recommending it, I'm just saying you can use it.

Enchanted Bow (2). Yes I'm serious. But no, its not that good. If for some reason you're using this as your main weapon, i mean, you can run away, attack it, and ignore aloof, with no risk of the attack trigger, because the trigger only cares at it's location. I think this is inefficient for Kohaku, but it is worth pointing out because it solves 2 problems. But it is slot expensive taking up some slots you'd much rather use for something else. There's also no synergy with blurse. I'm not even listing the stats. 3 slots. thats the cost.

Final Impressions. Kohaku has many tools even before seeing the entire Hemlock Vale box. Any of Spectral Razor, Drain Essence, Manipulate Destiny or even the Occult Lexicon I could easily see as staples for Kohaku whether he is main fighter, cluever, or flex. For my money the best path for the lowest risk is Manipulate Destiny. This costs 2 xp, but it's searchable, the cheapest option, and its pretty reliable to trigger if you already do curses, even if you haven't already played a Favor of the Moon, or your signature yet. And even if you aren't killing the Yurei, depending on the timing... just ignore it and run? It's really not terribly threatening. Also, if you need to kill a Renfield or an Initiate, maybe just take the test?

Looks like it will become "Bugging Yurei" in the future. Its unique mechanism that attacks investigator during chaos token revealing can cause plenty of issues.

Let me first: what if someone plays Hypnotic Gaze to cancel its attack during a skill test? Let's say the investigator is testing (3) and reveals a , then they plays Hypnotic Gaze to cancel that attack and then to reveal a chaos token following Hypnotic Gaze's ability. Before this revealing, should they put the token drawn before back to the chaos bag first? Though the rule specified that skill tests cannot be embedded, this chaos token revealing is actually not a skill test.