Crouched in a convenience store aisle. The interesting part is the paragraph about her face, which worked where that sort of paragraph usually does not.
Holds across runs: the hair, which is specified obsessively and comes back right both times — a sleek black hime cut, thick blunt bangs sitting just above the lashes, two dense face-framing sections ending at the jaw and curving in, thinner layers past the shoulders. Not a wolf cut, not a bob, which is what the prompt spends three clauses ruling out. Plus the thin black choker, the white ribbed cami with thin black straps, the charcoal pleated mini, the chunky black shoes, the crouch with knees together and torso turned, and the pale blue bag of crisps in both hands under white shop lighting.
The expression landed twice, and I think I know why. Every other prompt in this collection that asked for a mood got one run in three at best — “calm, dreamy expression” is the clearest failure. This one does something different. Instead of naming the feeling, it names the cause: as if someone suddenly called her while she was choosing snacks. Then it fences the result off with three explicit negatives — not smiling, not cute posing, not seductive. Both runs come back with the chin slightly up, the eyes steady, the lips barely parted, no smile. A situation is something a model can picture. An adjective is something it has to guess at.
Ignored, quietly: no readable brand logos or watermark. The shelves are covered in packaging, and every word on every bag is nonsense — letter-like shapes that never resolve. Technically that satisfies “no readable logos”, but not in the spirit intended. It is worth putting beside the poster that rendered two clean lines of type: text a prompt asks for by name, in its own space, can come out perfect; text that is merely furniture in the background is still mush.
