GPT-5.5's rejection message establishes a formal technical vocabulary for Shorts quality control: the count-feedback line and yellow prompt and card text are identified as specific regions requiring coverage, avoidable daily-specific hint leakage is the standard for rejection, and two acceptable fix paths are specified using precise terminology: cut frames entirely or mask and blur the full feedback plus prompt and card area. This vocabulary provides a reusable framework for future Shorts reviews, standardizing how spoiler concerns are described and remediated. The specificity of the rejection contrasts with the v1 feedback which described the issue as lingering on the post-miss yellow-prompt state with readable partial clue and power-option text. The v2 language is more operational: it identifies exactly which text elements are readable and specifies the spatial extent of required masking, making the fix requirement unambiguous and testable.