style-biblelisted
Install: claude install-skill popnow-ai/popnow-claude-plugin
# Style bible
Inconsistent AI visuals usually come from prompts that describe the style differently each time. A style bible fixes the vocabulary once and reuses it everywhere.
## Inputs
Brand name or campaign, audience, 3 adjectives for the feeling, any existing brand colors, fonts and references the user owns or likes (described in words), and the formats it will be used for.
## Sections
1. **Mood**: 3 adjectives and one sentence on the feeling.
2. **Palette**: 4 to 6 colors with hex codes and their role (background, accent, skin-tone-safe light, shadow). Note how they appear in scenes ("a terracotta wall", "mustard knitwear"), because models follow objects better than abstract colors.
3. **Light**: time of day, quality (soft, hard), direction, contrast, color temperature.
4. **Camera**: lens feel (24mm wide, 50mm natural, 85mm portrait), depth of field, movement vocabulary (slow push-in, locked-off, handheld), frame rate feel.
5. **Texture and medium**: film grain, clean digital, clay, watercolor, 3D render, and so on.
6. **Sets and props**: recurring locations and objects.
7. **People and wardrobe**: original characters only, described by clothing, hair and accessories, not by resemblance to real people.
8. **Overlay typography**: font families, sizes for 9:16 and 16:9, caption style. These go in the edit, not in generated frames.
9. **Prompt suffix**: one 25 to 40 word line that captures sections 2 to 5, to append to every image and video prompt.
10. **Do-not list**