WILD CORPUS · github_awesome
PQS 55 (C) - prompt from raw.githubusercontent.com
Source: raw.githubusercontent.com · Scraped 2026-05-04 · Scored 2026-05-04
Score
C55 / 80
gemma4:latest · local · pqs-v2.0 · canonical
Clarity9 / 10
Specificity10 / 10
Context10 / 10
Constraints9 / 10
Output format7 / 10
Role definition8 / 10
Examples1 / 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) involved in a covert exchange on a foggy train platform in 1940s London. The image must be photorealistic, featuring cinematic lighting and highly detailed textures of wool and steam. The aesthetic should look like it was shot on Arri Alexa with a cinematic depth of field, capturing the tension and romance of a noir thriller.",
"details": {
"year": "1944",
"genre": "Cinematic Photorealism",
"location": "A dimly lit, steam-filled railway platform in London at night, with the blurred silhouette of a locomotive in the background.",
"lighting": [
"Dramatic chiaroscuro",
"Volumetric lighting through steam",
"Cold atmospheric backlight",
"Warm tungsten practical light from a station lamp"
],
"camera_angle": "Over-the-shoulder close-up shot, focusing on the faces and the subtle hand exchange.",
"emotion": [
"Secretive",
"Urgent",
"Melancholic",
"Tense"
],
"color_palette": [
"Steel blue",
"Charcoal grey",
"Sepia highlights",
"Deep crimson"
],
"atmosphere": [
"Noir",
"Mysterious",
"Cinematic",
"Foggy"
],
"environmental_elements": "Thick billowing steam from the train engine, wet cobblestones reflecting light, vintage leather suitcases in the periphery.",
"subject1": {
"costume": "A textured heavy wool trench coat, a fedora hat slightly tipped forward, and leather gloves.",
"subject_expression": "Stoic and alert, eyes darting to the side to check for surveillance.",
"subject_action": "Discreetly slipping a small, sealed envelope into Subject 2's hand."
},
"negative_prompt": {
"exclude_visuals": [
"bright daylight",
"modern technology",
"smartphones",
"digital watches",
"modern architecture"
],
"exclude_styles": [
"cartoon",
"3d render",
"anime",
"oil painting",
"sketch",
"low resolution"
],
"exclude_colors": [
"neon green",
"fluorescent pink",
"oversaturated colors"
],
"exclude_objects": [
"cars",
"airplanes",
"plastic"
]
},
"subject2": {
"costume": "A tailored 1940s skirt suit with a fur collar, a pillbox hat with a small mesh veil, and red lipstick.",
"subject_expression": "Anxious but composed, biting her lip slightly, looking intently at Subject 1.",
"subject_action": "Grasping Subject 1's hand tightly while receiving the envelope, pulling her coat closer."
}
}
}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.