v1.2
- Added brief-adaptation cues and conditional background rules
- Documented per-section framing, composition variety, hero scale presets, conversion-oriented section roles, LAYOUT_VARIATION and CONVERSION_DISCIPLINE dials, and expanded clarity checks
- Added concept spine, second-read cues, and further anti-slop bans
- Tightened hero composition rules with a composition bias rather than a hard ban on default left-text or right-image hero layouts
- Required one horizontal reference image per section, disallowing multiple sections to be combined in one frame
- Expanded background modes and gradient discipline to encourage palette-matched gradients and discourage typical AI-default gradient recipes
imagegen-frontend-web: tightened hero composition rules (default left-text or right-image hero banned, then revised to a composition bias rather than a hard ban in a follow-up commit). imagegen-frontend-web: required one horizontal reference image per section. A page with eight sections expects eight images; combining multiple sections in one frame is not allowed. imagegen-frontend-web: added brief-adaptation cues plus conditional background rules. imagegen-frontend-web: expanded background modes and gradient discipline (palette-matched gradients encouraged, typical AI-default gradient recipes discouraged). imagegen-frontend-web: documented per-section framing, composition variety, hero scale presets, conversion-oriented section roles, LAYOUT_VARIATION and CONVERSION_DISCIPLINE dials, and expanded clarity checks. imagegen-frontend-web: concept spine, second-read cues, and further anti-slop bans.