We need to output JSON with fields: joke_text, headline, meme_caption, image_pro...

We need to output JSON with fields: joke_text, headline, meme_caption, image_prompt, meta_title, meta_description. Must be in German language for joke_text, headline, meme_caption, but image_prompt must be English. Must avoid repetition, be unique. Must start with "Chuck Norris" or mention him. Must be funny, classic style. Must be unique, not similar to existing jokes. Provide a new joke. We need to ensure JSON format, no markdown code block. Just raw JSON. We need to keep each field within length constraints: headline max 80 characters, meme_caption max 80 characters, meta_title max 60 characters, meta_description max 160 characters. Let's craft a joke: "Chuck Norris kann mit einem einzigen Blick das Internet zum Stillstand bringen, weil er die Daten schneller verarbeitet als ein Server-Farm-Cluster." That's a bit long but okay. But need to start with Chuck Norris. Maybe: "Chuck Norris hat einmal einen Kühlschrank so kalt gemacht, dass er die globale Erwärmung um 5 Grad senkte." That's a classic. But need to be unique, not repeating known jokes. Let's think of a new one: "Chuck Norris hat einmal einen Schneemann gebaut, der so stark war, dass er die globale Erwärmung um 10 Grad verringern konnte." That's similar to previous but maybe okay. Better: "Chuck Norris kann mit einem einzigen Niesen die Schwerkraft auf 0,5g reduzieren, sodass er im Weltraum tanzen kann." That's unique. Let's craft: "Chuck Norris hat einmal einen Kaugummi gekaut, der so stark war, dass er die Gesetze der Physik neu schrieb." That's okay. We need to ensure it's unique and not similar to known jokes. Let's produce something like: "Chuck Norris kann mit einem einzigen Handschlag einen Tornado in eine Kaffeepause verwandeln." That's unique. Now we need to fill fields. joke_text: full joke text in German, starting with Chuck Norris. headline: short headline max 80 chars. Something like "Chuck Norris lässt die Schwerkraft zittern". That's 33 chars. meme_caption: max 80 chars, maybe "Wenn Chuck Norris lächelt, schmilzt der Schnee". That's 44 chars. image_prompt: English prompt for image generation. Must describe scene from joke. For our joke about handshake turning tornado into coffee break. Let's craft: "A hyper-realistic portrait of a tough, legendary action hero man in his 50s with reddish-brown hair, thick reddish beard, piercing eyes, muscular build. Black sleeveless vest. Chuck Norris delivering a powerful handshake that transforms a swirling tornado into a calm coffee break scene, with coffee cups floating. Photorealistic action movie poster, dramatic cinematic lighting, American flag background. No text, no captions." Make sure it's English, includes [Spezifische Szene aus dem Witz] placeholder? The instruction says: "[Spezifische Szene aus dem Witz]" in the prompt. So we need to include that placeholder? The example includes that placeholder. Probably we need to include that exact phrase? The instruction: "A hyper-realistic portrait ... [Spezifische Szene aus dem Witz]. Photorealistic action movie poster..." So we need to embed the specific scene description where that placeholder is. So we should replace [Spezifische Szene aus dem Witz] with our scene description. So the prompt should contain that placeholder? Actually they want the prompt to include that placeholder? The example: "A hyper-realistic portrait ... [Spezifische Szene aus dem Witz]. Photorealistic action movie poster..." So we need to include that placeholder exactly? Probably yes, they want that placeholder to be included as part of the prompt, indicating where the specific scene goes. So we should include "[Spezifische Szene aus dem Witz]" exactly in the prompt. Then after that, continue with the rest. So the prompt should be something like: "A hyper-realistic portrait of a tough, legendary action hero man in his 50s with reddish-brown hair, thick reddish beard, piercing eyes, muscular build. Black sleeveless vest. [Spezifische Szene aus dem Witz] Photorealistic action movie poster, dramatic cinematic lighting, American flag background. No text, no captions." But we need to fill the scene: maybe "Chuck Norris performs a handshake that turns a tornado into a coffee break". So we can put that as the scene description. So the prompt becomes: "A hyper-realistic portrait of a tough, legendary action hero man in his 50s with reddish-brown hair, thick reddish beard, piercing eyes, muscular build. Black sleeveless vest. Chuck Norris performs a handshake that turns a tornado into a coffee break Photorealistic action movie poster, dramatic cinematic lighting, American flag background. No text, no captions." But we need to keep the placeholder? The instruction says: "[Spezifische Szene aus dem Witz]" is part of the prompt. So we should include that exact token, not replace it. Actually they want the prompt to contain that placeholder, which we replace with the specific scene. The example shows that placeholder is included, then after that they continue with "Photorealistic action movie poster...". So we should include the placeholder as is, then after that we add the scene? Or we replace the placeholder with the scene? The instruction: "A hyper-realistic portrait of a tough, legendary action hero man in his 50s with reddish-brown hair, thick reddish beard, piercing eyes, muscular build. Black sleeveless vest. [Spezifische Szene aus dem Witz]. Photorealistic action movie poster, dramatic cinematic lighting, American flag background. No text, no captions." Thus we need to put the placeholder where the scene description goes, then close the bracket? Actually they have "[Spezifische Szene aus dem Witz]" inside brackets. So we should replace that with our specific scene description, but keep the brackets? Probably we should replace the placeholder with the actual scene description, but keep the brackets? The instruction says "Englischer Prompt für Bildgenerierung: A hyper-realistic portrait ... [Spezifische Szene aus dem Witz]. Photorealistic action movie poster..." So the placeholder is meant to be replaced with the specific scene from the joke. So we should put our scene description there, without brackets. So the final prompt should be: "A hyper-realistic portrait of a tough, legendary action hero man in his 50s with reddish-brown hair, thick reddish beard, piercing eyes, muscular build. Black sleeveless vest. Chuck Norris performs a handshake that turns a tornado into a coffee break. Photorealistic action movie poster, dramatic cinematic lighting, American flag background. No text, no captions." That seems fine. Now