WILD CORPUS · github_awesome
PQS 60 (B) - prompt from raw.githubusercontent.com
Source: raw.githubusercontent.com · Scraped 2026-05-04 · Scored 2026-05-04
Score
B60 / 80
gemma4:latest · local · pqs-v2.0 · canonical
Clarity10 / 10
Specificity10 / 10
Context10 / 10
Constraints10 / 10
Output format8 / 10
Role definition9 / 10
Examples2 / 10
CoT structure1 / 10
The prompt
{
"prompt": "You will perform an image edit using the people from the provided photos as the main subjects. Preserve their core likeness. Create an Ultra-Photorealistic, Movie-Quality scene depicting Subject 1 (male) and Subject 2 (female) as covert spies meeting on a foggy, iron bridge during the Cold War. The image must look like a frame from a high-budget blockbuster movie shot on Arri Alexa. Use cinematic lighting to create deep shadows and highlights. The scene is highly detailed with a shallow depth of field. Subject 1 is handing off a secret package to Subject 2. The composition adheres to a cinematic 1:1 aspect ratio.",
"details": {
"year": "1962",
"genre": "Cinematic Photorealism",
"location": "The Glienicke Bridge at midnight, obscured by thick river fog and illuminated by dim, yellow streetlamps.",
"lighting": [
"Volumetric fog lighting",
"Noir style chiaroscuro",
"Rim lighting on silhouettes",
"Soft yellow tungsten glow"
],
"camera_angle": "Eye-level medium closeup with a shallow depth of field to isolate the subjects from the misty background.",
"emotion": [
"Suspenseful",
"Urgent",
"Clandestine"
],
"color_palette": [
"Steel blue",
"Fog gray",
"Tungsten amber",
"Deep black",
"Vibrant crimson"
],
"atmosphere": [
"Cold",
"Tense",
"Cinematic",
"Mysterious"
],
"environmental_elements": "Swirling mist rising from the water below, damp iron railings, the distant blurred headlights of a vintage checkpoint vehicle.",
"subject1": {
"costume": "A textured charcoal wool peacoat with the collar turned up against the wind.",
"subject_expression": "Anxious, with sweat glistening on his brow and eyes darting nervously.",
"subject_action": "Subtly sliding a leather dossier across the railing towards Subject 2."
},
"negative_prompt": {
"exclude_visuals": [
"daylight",
"sunshine",
"modern cars",
"cell phones",
"neon signs"
],
"exclude_styles": [
"cartoon",
"3D render",
"sketch",
"anime",
"impressionist"
],
"exclude_colors": [
"neon green",
"hot pink",
"pastel colors"
],
"exclude_objects": [
"umbrellas",
"crowds",
"modern architecture"
]
},
"subject2": {
"costume": "A classic beige trench coat belted at the waist and a red hat.",
"subject_expression": "Stoic and composed, with a piercing, calculating gaze.",
"subject_action": "Reaching out with a black leather-gloved hand to intercept the dossier while looking over her shoulder."
}
}
}This prompt was scraped from a public source. The score reflects the input as written, not the quality of any output it produced. The AI input quality problem is the gap between what people type and what the model can act on.