T-Shirt Design Engine Prompt (GPT 2.5)

2x2 grid, 16:9, 4 topics that are relevant today: T₁ COMBINATORIAL FUSION (The 10×2 Law) 10 categories, 2 distinct picks each (20 elements total). Fused via CHANNEL OWNERSHIP: one style owns typography, one owns color, one owns composition, one owns texture. Conflicts are resolved by dominance, not stacked. ¬default to Bauhaus unless explicitly rolled. ¬T₁ = random collage / Bauhaus default / style stacking. T₂ SUBJECT ABSTRACTION (The Wearable Icon) The subject is transformed via the selected Image Transformation System (e.g., diagrammed, mascot-ified, emblem-simplified). It retains its core identity/silhouette but becomes a graphic icon, NOT a literal photograph or boring clip-art. ¬T₂ = literal photo / unrecognizable / boring stock vector. T₃ PRINT PHYSICS (The Screen-Print Law) Limited color count (apparel-safe), print-aware geometry (no micro-details that bleed on fabric), authentic print texture (riso, plastisol, puff, halftone). ¬T₃ = muddy gradients / infinite colors / micro-clutter. T₄ TYPOGRAPHIC DISCIPLINE (The Merch Text Law) Short, punchy, graphic headline + optional micro-copy. Style-matched font. ZERO long paragraphs or fake lorem ipsum. Readable on the human torso. ¬T₄ = fake paragraphs / unreadable text / wrong font. T₅ MERCHANDISE COMPOSITION (The Product Shot Law) Designed for the human form (chest emblem, oversized front, spine). Strong contrast, clear silhouette, professional mockup/finish. Looks like something people would actually buy. ¬T₅ = bad mockup / warped graphic / floating design.