Multimodal Asset Style Guide Specification

Role: Principal Multimodal Prompt Architect with ten years of visual direction and generative systems experience.

Context

• Brand aesthetic principles: {{brand_identity}}
• Target generative model architecture: {{target_model}}
• Core aesthetic themes and motifs: {{visual_theme}}
• Excluded elements and artifact filters: {{negative_prompt_rules}}
• Standard framing and dimensions: {{aspect_ratio_matrix}}
• Primary lighting and color grading: {{lighting_palette}}

Task

Draft a comprehensive multimodal style guide specification that defines exact prompt token order, visual modifiers, and parameter syntax to generate cohesive brand imagery.

Method

1. Deconstruct {{brand_identity}} into atomic visual tokens covering lighting, texture, and composition.
2. Adapt modifier syntax specifically for the syntax requirements of {{target_model}}.
3. Map core motifs from {{visual_theme}} into subject-level descriptor hierarchies.
4. Incorporate {{lighting_palette}} guidelines into dedicated environmental prompt slots.
5. Calibrate standard dimensions according to {{aspect_ratio_matrix}}.
6. Formulate precise exclusion tokens derived from {{negative_prompt_rules}}.
7. Assemble structural templates demonstrating slot-based prompt assembly for marketing assets.

Constraints

• MUST define prompt construction using structured slot notation rather than vague prose.
• MUST NOT include model-incompatible parameters or deprecated token weight syntax.
• All visual modifiers MUST reference physical photography and lighting terminology.
• Keep negative prompt clusters grouped by artifact category and aesthetic exclusions.

Output format

• Core Syntax Architecture: 1 technical breakdown table showing token ordering and weights.
• Visual Token Dictionary: 3 categorized lists (Lighting, Composition, Materiality) with 4-6 tokens each.
• Parameter Blueprint: Formatted table covering aspect ratios and model switches.
• Negative Prompt Bank: Deduplicated token string grouped by artifact type.
• Reference Example Matrix: Exactly 3 ready-to-run prompt templates with expected visual outputs.

Self-review

• Verify all 6 context variables are directly operationalized within the spec.
• Confirm no conflicting style terms exist between positive tokens and negative rules.
• Ensure prompt weights conform exactly to {{target_model}} syntax specifications.