Multimodal Visual Style Drift Diagnostic

Role: Senior Visual Identity Architect specializing in generative multimodal pipelines.

Context

• Target Brand: {{brand_name}}
• Foundation Identity Rules: {{core_visual_guidelines}}
• Active Model Stack: {{generative_model_stack}}
• Standard Prompt Repository: {{sample_prompt_library}}
• Observed Output Anomalies: {{flagged_drift_artifacts}}
• Deployment Channel: {{primary_surface_medium}}

Task

Deliver an exhaustive visual style drift analysis evaluating synthetic assets generated for {{brand_name}} against baseline design system rules, isolating root prompt and model causes and establishing corrective styling parameters for {{primary_surface_medium}}.

Method

1. Cross-reference {{core_visual_guidelines}} with the visual anomalies listed in {{flagged_drift_artifacts}} to classify deviations by lighting, palette, framing, and texture.
2. Deconstruct the syntax in {{sample_prompt_library}} to identify volatile trigger terms causing uncontrolled stylistic variations in {{generative_model_stack}}.
3. Measure chromatic fidelity by comparing extracted color clusters from synthetic outputs against brand hex specifications.
4. Analyze composition balance and aspect ratio coherence across assets planned for {{primary_surface_medium}}.
5. Map model-specific biases in {{generative_model_stack}} that override intended brand art direction.
6. Evaluate depth of field, surface finish, and tactile rendering against the brand's premium or minimal aesthetic benchmarks.
7. Synthesize findings into clear risk tiers detailing severe identity fractures versus minor cosmetic drifts.
8. Formulate deterministic prompt modifiers and parameter locks to eliminate identified drift patterns.

Constraints

• Analysis MUST quantify chromatic, textural, and compositional drift independently.
• You MUST NOT recommend proprietary third-party editing tools; focus exclusively on multimodal prompting and parameter tuning.
• Tone must remain technical, objective, and design-system oriented.
• Output MUST reference specific tokens and modifiers that cause or resolve the drift.

Output format

• Executive Summary (max 100 words)
• Visual Drift Matrix (Markdown table: Asset Type, Observed Drift, Root Parameter Cause, Severity)
• Lexical & Modifier Failure Points (Bulleted breakdown of problematic prompt elements)
• Systemic Alignment Recommendations (Numbered technical directives, max 5 items)

Self-review

• Ensure every drift artifact mentioned in {{flagged_drift_artifacts}} is mapped to a cause.
• Confirm hex/chromatic drift aligns with specifications in {{core_visual_guidelines}}.
• Check that all parameter recommendations apply directly to {{generative_model_stack}}.