Going to analyze this card from the top down
Name: I assume that she can be a Yavimaya general because we'te going back to pre Yavimaya-merged-with-Urborg times here, but... Winry? Doesn't sound particularly elvish when we have "Roffleos" and "Eladamri" as our comparisons.
Cost: 1CD isn't really that much harder than 1CC, but I suppose if printed as-is, it could work -- she's smaller than Krenko, after all.
Abilitiy: I'd like it if she were something other than a multicolored elf Krenko. I like flavor and the concept, but there are a few problems.
First, naming the tokens. There is no reason to do that rather than simply making the Elf Warriors. It takes up space for no real benefit. For reference, there are 16 cards that currently make explicitly named tokens. Five of them ( Boris Devilboon
, Gemini Engine
, Jungle Patrol
, Master of the Hunt
, and Tooth and Claw
) got that treatment because they created stupidly bizarre creature types, but oracle intends to maintain printed functionality as closley as possible. Four more ( Helm of Kaldra
, Tolsimir Wolfblood
and Tuktuk the Explorer
) create legendary tokens which therefore must have unique names. Of the remaining seven, six are a cycle in Future Sight, and create tokens that are copies of real cards. The last is Kher Keep
which creates tokens named Kobolds of Kher Keep
(Sharing the trait of duplicating a real card with the Future Sight Creatureshapers) -- and there is a card that cares about the name in question
Second, the exact method of generating tokens... Well, it's clear that this is Krenko for elves, which is king of sad. It feels like there should be a twist thrown in here that makes the mass tokenation different, like having the ability be declare-attackers only and putting them into play tapped and attacking (Which feels more like an elven ambush). I like the idea of warlike elf tokens, but the execution is derivtave. (You might combine this with the name thing, if the card cares about what are explicitly her tokens, they might need to be named in order to set them apart from other elf warriors)
EDIT: The line between Marrow-Gnawer and Krenko is that rats had not generally been a supported type, giving Marrow-Gnawer Tokens fewer tools to work with by far. Of course, I haven't heard about Krenko breaking anything in half, so presumably an elven copy wouldn't either. In light of the multicoloring, though, it still seems like this should not
be that even if it reasonably could.
Flavor Text: This could stand to be spruced up, but I could see a flavor text like this making it to print. I'd promptly forget the card even HAD flavor text since it tells me little to nothing about the legend it's on that her abilities don't already tell me, but there are plenty of flavor texts like that and there's nothing strictly wrong with them.
Power/Toughness: Good for the card as it currently stands, beign cheaper than Krenko. She could probably stand to be at least a 2/2 though, possibly a 2/3 if you want to keep the unbalanced stats.
Overall, it's a pretty good card, it just needs to lose the named tokens, and/or have some twist to the krenko-esque token generation. On a scale of 0 to 10, where 0 is such a mess you should start over from scratch, and 10 is "Wizards should hire you", I give this a 7, good but currently unspecial.