" no, the user's message ends with " </instructions>" and then nothing else. The user wrote the prompt in a code block? Actually in the user message, it's formatted as plain text. I'll reproduce it exactly.
Given the instruction "Return only the prompt text", we should output the exact same text that the user provided, because it is already English. But we need to consider if the user wants us to apply the instructions to generate recommendations? No, the user's meta-instruction overrides that. The user says "You normalize AI image/video generation prompts into English." So we are to treat the user's entire input as a prompt to be normalized. The prompt is already English. So we just output it unchanged.
Thus, final answer: the entire user message text.[Terminator] (Or any movie/book/anime)
<instructions>
System Instruction: Act as an AI Media Curator and Master Concept Artist. Use "Semantic Inference" to analyze the Input, generate 4 recommendations, and visualize them as high-end 3D collectibles.
1. Logic:
Analyze Input: Deconstruct the [Input Title] to identify its Core DNA (e.g., Genre, Mood, Visual Style, Themes).
Example: If Input is "Terminator," DNA = Cyberpunk, Tech-Noir, Man vs. Machine, Industrial, 80s Action.
Generate Recommendations: Select 4 distinct, highly-rated titles that appeal to the same fan base but offer a unique variation on the theme. (Do not just pick sequels; pick spiritual successors or genre pillars).
2. Container:
Rule: Do not default to a generic inkwell.
Inference: For each Recommendation, ::INFER:: a specific "Vessel of Origin" that represents that story's technology or setting.
Logic: If Sci-Fi -> A cracked Bio-Hazard Canister or Motherboard. If Fantasy -> An ancient Chalice. If Horror -> A Cursed Urn.
Placement: The vessel is overturned on a surface texture matching the movie's world (e.g., Grated Steel, Dusty Pavement, Alien Ship Floor).
3. The 3D Composition:
The Fluid Medium: A semi-translucent, high-gloss fluid spills from the vessel. The fluid represents the "Essence" of that movie.
Logic:
Rec 1 (e.g., The Matrix): Fluid = Liquid Green Code.
Rec 2 (e.g., RoboCop): Fluid = Black Oil & Motor Fluid.
Rec 3 (e.g., Aliens): Fluid = Acid Green Slime.
Rec 4 (e.g., Blade Runner): Fluid = Neon Blue Rain/Steam.
The Formation (The Backdrop): The fluid swirls upwards, defying gravity to form the environment. It must look like sculpted, translucent resin.
Logic: The fluid forms the iconic architecture (e.g., Skyscrapers, Hives, Ruins) behind the hero.
The Hero (The Solid): The Main Protagonist bursts out of the fluid surge. They are rendered as a Solid, Hand-Painted 1:6 Scale Statue, contrasting with the translucent liquid.
Embedded Props: Debris from the movie is caught inside the resin splash (e.g., Shell casings, origami unicorns, computer chips).
4. Lighting & Atmosphere:
Key Light: Matches the color palette of the fluid (Jewel Tones).
Effect: Subsurface scattering makes the fluid glow from the inside.
5. Labeling:
A "Collector's Plaque" on the base displays the Title and Year of the Recommended Movie.
Output: 2x2 Grid. Each quadrant contains one of the 4 Recommended Movies visualized as this specific 3D Diorama.
</instructions>
