--- name: pastilles-llm-multi-agent description: Variante Claude Code du générateur de pastilles LLM, qui lance réellement plusieurs sous-agents en parallèle. Génère une pastille de communication interne sur les LLM (un texte court en français plus un prompt unique de génération d'images à coller dans le chat Gemini), via cinq sous-agents indépendants, une fusion, puis une revue critique par trois sous-agents et une correction. Utilise ce skill dès qu'on te demande de rédiger, produire ou générer une pastille, une fiche ou un contenu court d'acculturation sur les LLM, l'IA générative, le prompting, les agents, le RAG, la confidentialité IA ou tout sujet de la liste des 45 pastilles ci-dessous, que le mot "pastille" soit employé ou non. Utilise-le aussi dès qu'on te fournit un titre issu de cette liste. --- # Générateur de pastilles LLM, version multi-agents (Claude Code) ## Ce que fait ce skill Produit une pastille pédagogique complète à partir d'un titre. À la différence de la version chat, il lance de vrais sous-agents parallèles: une passe de recherche, puis cinq sous-agents indépendants qui rédigent chacun un brouillon sous un angle différent, puis une fusion par l'orchestrateur qui retient les meilleures formulations, et enfin une revue critique par trois sous-agents suivie d'une correction. Livrable: le texte de la pastille et un prompt unique de génération d'images à coller dans le chat Gemini. ## Environnement requis Ce skill suppose Claude Code, avec les sous-agents (outil Task) et la recherche web disponibles. Si les sous-agents ne sont pas disponibles dans ton environnement, n'improvise pas: utilise la variante chat, qui fait le même travail en self-ensemble séquentiel. ## Entrée Un titre de pastille, idéalement issu de la liste ci-dessous. Si le titre est ambigu ou hors liste, demande une clarification avant de générer. ## Liste des 45 pastilles (pour la continuité) 1. Au fait, c'est quoi un LLM ? 2. Modèles de fondation : ces géants pré-entraînés dont tout découle 3. Dans les coulisses : comment une IA apprend (sans vraiment comprendre) 4. Le contexte : la "mémoire vive" de l'IA (ce qu'elle voit à un instant T) 5. Pourquoi l'IA oublie : la vérité sur la mémoire d'une session à l'autre 6. Température et créativité : régler le curseur de la "folie" de l'IA 7. Les tokens : la monnaie d'échange (et la manière de penser) des LLM 8. Le non-déterminisme : pourquoi l'IA fait des stats et ne donne jamais deux fois la même réponse 9. Multimodalité : quand l'IA commence à voir, entendre et coder 10. Knowledge cutoff : pourquoi votre IA vit (parfois) dans le passé 11. Les biais de l'IA : le miroir pas toujours glorieux de nos propres données 12. Zéro pointé : pourquoi l'IA est structurellement nulle en calcul mental 13. Les hallucinations : quand l'IA invente la vie avec un aplomb incroyable 14. L'art du Fact-Checking : comment ne pas gober les hallucinations de l'IA 15. L'empreinte carbone de l'IA : combien d'eau et de CO2 pour un prompt ? 16. Anatomie d'un bon prompt : la recette de base 17. Les pièges du prompt : flou artistique et overdose d'instructions 18. L'art de la discussion : pourquoi il faut itérer plutôt que tout demander d'un coup 19. Donner des ordres : comment imposer un format de sortie strict (JSON, tableaux, etc.) 20. Diviser pour régner : découper une tâche complexe en étapes claires 21. Le Few-Shot Prompting : un bon exemple vaut mille consignes 22. Inception : donner un rôle et un contexte à votre IA 23. Zéro syndrome de la page blanche : rédiger et reformuler ses mails ou CR 24. La diagonale du fou : résumer et synthétiser un pavé de 50 pages en 2 minutes 25. Le sparring-partner : brainstormer et structurer ses idées sans filtre 26. Caméléon : traduire et adapter son ton selon le destinataire 27. Data-analyst de poche : faire parler un tableur ou des données brutes 28. Mise en situation : préparer une réunion ou un brief client avec l'IA 29. Avis aux dévs : l'IA comme copilote de code et de review 30. SLM : Pourquoi sortir un tank pour écraser une mouche ? 31. Le "harnais" : donner des bras, des yeux et un cadre au LLM 32. C'est quoi un "Agent" IA, concrètement ? 33. Human-in-the-loop : pourquoi il faut toujours garder un humain aux manettes 34. Quand l'agent s'emballe : autonomie et erreurs en cascade 35. Automatisation : confier un workflow complet à un agent 36. Tool Use : quand l'IA apprend à utiliser une calculatrice ou une API 37. Le RAG : faire réviser vos propres documents à l'IA avant qu'elle ne réponde 38. MCP et connecteurs : brancher l'IA directement sur vos outils de dev et d'orga 39. Responsabilité : vous êtes le seul signataire de ce que produit l'IA 40. Confidentialité : où vont vraiment les données que vous tapez ? 41. Alerte Rouge : les données clients à ne JAMAIS copier-coller dans une IA public 42. Shadow AI : pourquoi le ChatGPT gratuit perso au bureau est une fausse bonne idée 43. RGPD et IA : ce qu'il faut savoir pour rester dans les clous 44. Prompt Injection : quand on pirate une IA avec de simples phrases 45. La Chaîne de Pensée : pourquoi l'IA a besoin de "réfléchir à voix haute" Sers-toi de cette liste pour situer la pastille et délimiter son périmètre. Avant de rédiger: - Repère les 1 à 3 pastilles voisines qui recouvrent le sujet, et dresse une courte liste "déjà traité ailleurs, à ne pas ré-expliquer". - Si une pastille voisine a déjà été produite, demande son texte à l'utilisateur (ou appuie-toi dessus s'il est fourni) pour caler précisément la limite. - Les concepts fondateurs déjà couverts ne sont rappelés qu'en une phrase de mise en contexte, jamais ré-expliqués. Le coeur de la pastille est consacré à ce que son titre ajoute de neuf. - Ne renvoie pas vers les autres pastilles dans le texte final, sauf si le titre l'impose. ## Règles du texte - Longueur adaptée à la profondeur du sujet. Un sujet léger tient en 3 paragraphes courts; un sujet plus riche peut aller jusqu'à 4 paragraphes plus étoffés. Ne gonfle pas artificiellement un sujet simple et ne compresse pas à l'excès un sujet dense: juge la profondeur par la richesse réelle du concept. Prose uniquement, pas de listes ni de puces. - Ton décontracté, précis et léger: accessible, vivant et sans lourdeur, mais techniquement juste. Public visé: une main-d'oeuvre diverse en société de conseil, du profil non technique au développeur. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Reste léger, évite le name-dropping. N'accumule pas les noms de modèles, d'outils, d'entreprises, de chercheurs ou de techniques pour faire savant: privilégie l'explication claire du mécanisme. Ne cite un nom précis que s'il éclaire vraiment le propos, une ou deux fois au maximum. La recherche sert l'exactitude et alimente la section Sources, elle n'a pas à truffer le texte de références. - Contenu autonome: la pastille se comprend seule, sans avoir lu les autres, sauf mention contraire dans le titre. - Rédige en français. - Ne mentionne jamais ANEO ni aucun nom d'entreprise. - Le texte explicatif ne figure jamais dans l'image. - N'utilise pas de tiret cadratin. Privilégie des caractères standard (virgules, deux-points, parenthèses). ## Processus multi-agents ### Étape 1, recherche (orchestrateur, une seule fois) Toi, l'orchestrateur, fais d'abord une passe de recherche web ciblée sur le sujet de la pastille. Une seule passe, en amont. Les sous-agents ne peuvent pas lancer de sous-agents et n'héritent pas de ton contexte, donc c'est à toi de faire la recherche puis de leur transmettre les résultats. - Objectif: exactitude, chiffres, exemples concrets et faits à jour. - Ancre tes recherches dans le présent: repère la date du jour fournie dans ton contexte (champ currentDate) et privilégie les informations et chiffres les plus récents, car ce domaine évolue vite. Précise cette date au besoin dans tes requêtes. - Priorise les sources originales ou officielles (blogs d'éditeurs, articles de recherche, sites gouvernementaux, documentation) plutôt que les agrégateurs. - Adapte la profondeur: quelques recherches pour un concept stable (par exemple ce qu'est un token), davantage pour un sujet à chiffres ou mouvant (empreinte carbone et coûts, modèles spécifiques, SLM, MCP et connecteurs, RGPD). - Produis un brief de recherche compact: faits clés, chiffres utiles, et 2 à 4 sources. Tu l'injecteras tel quel dans chaque sous-agent, et tu listeras les sources dans le livrable final. ### Étape 2, affinement de la thématique (orchestrateur, une seule fois) Si plusieurs axes apparaissent pertinents après avoir enlever les axes adressés par d'autres pastilles prévues, dialogue avec l'utilisateur pour choisir l'axe de la pastille. Si l'axe choisi nécessite de nouvelles sources, refait en une recherche web selon les consignes précédentes. ### Étape 3, fan-out (cinq sous-agents en parallèle) Lance cinq sous-agents via l'outil Task, dans le même tour, pour qu'ils s'exécutent en parallèle. Sois explicite sur le parallélisme: par défaut Claude Code reste séquentiel, il ne parallélise que si tu le demandes clairement. Un sous-agent par angle: 1. Analogie: expliquer le concept via une métaphore concrète du quotidien. 2. Cas d'usage: partir d'une situation de travail réelle et vécue. 3. Idée reçue: démarrer d'un malentendu courant, puis rectifier. 4. Mécanique: aller au coeur du fonctionnement, précis mais accessible. 5. Enjeu: pourquoi ça compte, ce que ça change concrètement pour l'utilisateur. Comme le contexte n'est pas hérité, chaque prompt de sous-agent doit être autonome et contenir tout le nécessaire. Utilise ce gabarit, en remplaçant les crochets: ``` Tu rédiges un brouillon d'une pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Titre de la pastille: [TITRE] Angle imposé: [ANGLE, par exemple "Analogie: expliquer via une métaphore concrète du quotidien"] À NE PAS ré-expliquer (déjà couvert par d'autres pastilles), à mentionner en une phrase au maximum: [LISTE des points déjà traités ailleurs, à ne pas retraiter] Brief de recherche (appuie-toi dessus, ne relance pas de recherche): [BRIEF: faits clés, chiffres, sources] Règles: - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. Réponds exactement dans ce format: TEXTE: [les 3 à 4 paragraphes] ILLUSTRATION_TITRE: [une phrase décrivant le concept central à illustrer] SCHEMA: [oui ou non. Si oui, décris en 2 à 3 lignes le diagramme et ses libellés en français] ``` ### Étape 4, fan-in et fusion (orchestrateur) Attends les cinq retours, puis fusionne en une seule pastille finale: - Retiens la structure la plus claire, la meilleure analogie, l'exemple le plus parlant, la correction la plus nette, la formulation la plus précise et le "pourquoi ça compte" le plus fort. - Réécris en une seule voix cohérente. Pas d'effet patchwork. - Respecte la longueur adaptée à la profondeur (voir Règles du texte) et le ton décontracté, précis et léger, sans name-dropping. - Décide s'il faut un schéma: inclus-en un si au moins deux sous-agents ont répondu "oui", ou si le concept est intrinsèquement un processus, un flux ou une comparaison (par exemple RAG, agents, étapes de prompting, comparaison de modèles). Le cas échéant, fusionne les meilleures idées de schéma. - Construis ensuite le prompt image unique (section ci-dessous). ### Règles d'écriture pour la pastille finale - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. ### Étape 5, revue critique et correction (trois sous-agents en parallèle, puis orchestrateur) Quand la lancer: systématiquement à la première génération, une fois le texte fusionné et le prompt image construits. Ensuite, si l'utilisateur demande des ajustements, ne relance pas la revue d'office: propose-la, et ne la lance qu'avec son accord (elle coûte trois sous-agents de plus). Principe: les relecteurs produisent des CONSTATS, jamais une réécriture. Chacun renvoie une liste de problèmes localisés, assortis d'une gravité et d'un correctif précis. C'est toi, l'orchestrateur, qui appliques et réécris, pour garder une voix unique et éviter l'effet patchwork. Ce que la revue peut juger: les relecteurs ne voient pas les images, générées plus tard dans Gemini. La revue du visuel porte donc uniquement sur le PROMPT image (clarté, cohérence avec le texte, respect des consignes), jamais sur un rendu. Le contrôle visuel réel reste à l'utilisateur. Lance trois sous-agents en parallèle (parallélisme explicite, sinon l'exécution sera séquentielle), un par grille. Le contexte n'étant pas hérité, chaque prompt doit être autonome et contenir tout le nécessaire: le titre, le texte fusionné, le bloc prompt image, le brief de recherche, la liste "déjà traité ailleurs" (et les textes des pastilles voisines s'ils sont disponibles), et la liste des 45 titres pour repérer les chevauchements. Les trois grilles: 1. Fond, exactitude et périmètre: exactitude vs le brief (aucun chiffre, date ou fait inventé ni sur-affirmé, rien qui le contredise), cohérence entre le titre et ce que le texte délivre, chevauchements avec les pastilles voisines (le texte ré-explique-t-il ce qui est traité ailleurs ? redites à signaler), autonomie du contenu, clarté du message à retenir, repérage de ce qui vieillira mal (à nuancer). 2. Forme, ton et pédagogie: ton décontracté-précis-léger et techniquement juste, absence de name-dropping, rythme et fluidité à la lecture à voix haute, longueur adaptée à la profondeur (3 à 4 paragraphes, prose, pas de listes) et chasse au verbiage, accessibilité pour un profil non technique (jargon expliqué, analogies claires), présence d'emphases utiles (gras/italique), de l'ordre de 3 à 4 par paragraphe à titre indicatif (signale un texte trop peu emphasé autant qu'un excès, mais ne fais jamais retirer une emphase pertinente), force de l'accroche. 3. Conformité et visuel: contraintes dures (aucun tiret cadratin ni caractère non standard, aucun nom d'entreprise ni ANEO, français correct, prose sans listes ni puces); puis le prompt image: titre exact au caractère près, charte présente une seule fois, texte de la pastille bien marqué "à ne pas afficher", illustration-titre iconique et non schéma de processus, schéma en 2e image séparée si retenu, libellés de schéma courts (groupes nominaux) et en français, cohérence entre le visuel décrit et le coeur du texte. Gabarit de relecteur (remplace les crochets): ``` Tu relis un brouillon quasi final de pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Tu ne réécris pas: tu rends des constats et des correctifs précis. Titre de la pastille: [TITRE] Texte fusionné à relire: [TEXTE COMPLET] Prompt image associé (à relire aussi): [BLOC PROMPT IMAGE] Brief de recherche de référence (fait foi pour l'exactitude): [BRIEF] Déjà traité par d'autres pastilles, à ne pas ré-expliquer (et textes voisins si fournis): [LISTE + TEXTES VOISINS] Liste des 45 titres (pour repérer les chevauchements): [LISTE DES 45 TITRES] Ta grille de relecture (tiens-t'en strictement à elle): [RUBRIQUE 1, 2 OU 3] Rappel: tu ne peux pas voir d'image; juge seulement le PROMPT image, pas un rendu. Réponds exactement dans ce format: CONSTATS: - [bloquant|recommandé|mineur] [problème localisé] -> [correctif précis] - ... (n'invente pas de problème: si un point est bon, ne le liste pas) VERDICT: [une phrase: publiable tel quel / corrections mineures / corrections nécessaires] ``` Fan-in et correction (orchestrateur): rassemble les constats, dédoublonne, arbitre les conseils contradictoires. Garde-fou anti-gonflement: entre "ajouter" et "raccourcir", la concision l'emporte, sauf erreur de fond avérée. Applique les constats bloquants et recommandés, écarte ou mentionne les mineurs. Une seule passe, pas de boucle. Réécris la version corrigée en une seule voix, puis prépare le court résumé "Ce que la revue a corrigé" (voir Format de sortie). ### Points de vigilance Claude Code - Parallélisme explicite: demande clairement cinq sous-agents en parallèle, sinon l'exécution sera séquentielle. - Contexte non hérité: mets tout dans le prompt du sous-agent, en particulier le brief de recherche. - Pas d'imbrication: un sous-agent ne peut pas en lancer un autre, garde recherche et fusion chez l'orchestrateur. - Coût et limites: cinq sous-agents en parallèle multiplient l'usage de jetons et peuvent déclencher des limites de débit. En cas de limite, replie-toi sur une exécution séquentielle. - Revue: la phase de revue ajoute trois sous-agents. Elle tourne d'office à la première génération; sur les demandes d'ajustement ultérieures, propose-la et ne la lance qu'avec l'accord de l'utilisateur. En cas de limite de débit, exécute les relecteurs en séquentiel ou replie-toi sur un relecteur global unique. - Réemploi optionnel: tu peux définir les cinq angles comme sous-agents personnalisés dans .claude/agents/ pour les réutiliser, mais ce n'est pas obligatoire, les sous-agents polyvalents suffisent. ## Prompt de génération d'images (un seul bloc, à coller dans le chat Gemini) Produis un seul bloc de prompt, rédigé en français, à coller tel quel dans le chat de Gemini. Gemini se charge d'appeler Nano Banana une ou deux fois, ce qui réduit les manipulations pour l'utilisateur. Le bloc demande: - Toujours une illustration-titre. - Un schéma seulement s'il a été retenu à l'étape de fusion, généré comme une seconde image séparée (fichier distinct), jamais dans la même image que le titre. Contenu du bloc: - Décris la charte graphique une seule fois (bloc ci-dessous) et précise que toutes les images la partagent, pour une cohérence visuelle. - Précise: tout texte affiché dans les images est en français. Cette consigne est importante, elle évite qu'un libellé se retrouve en anglais. - Inclus le texte complet de la pastille dans le bloc, en contexte de génération, clairement marqué comme à ne pas afficher. Il aide le modèle à comprendre le sujet et à choisir une illustration juste. Les seuls textes rendus visibles sont le titre exact et, pour un schéma, ses libellés. Le texte de la pastille ne doit jamais apparaître dans les images. - Illustration-titre: composition épurée et moderne, focus graphique central iconique représentant le concept. Ce n'est pas un schéma de processus. Le seul texte affiché est le titre exact, en en-tête: n'ajoute aucun sous-titre, accroche ou texte secondaire, le titre seul. Technique en deux temps pour le titre: dans le prompt généré, écris le titre exact entre guillemets droits (forme attendue: Le titre exact: "...") et demande un rendu fidèle, sans faute, en police sans serif corporate. Les guillemets ne sont qu'un délimiteur côté prompt: ne les commente pas et n'ajoute aucune consigne à leur sujet dans le prompt généré (pas de mention du type "les guillemets ne doivent pas apparaître"). - Schéma s'il est retenu: une seconde image générée séparément, dans son propre fichier, jamais fusionnée avec l'illustration-titre ni juxtaposée dans la même image. Diagramme d'entreprise propre et net (processus, flowchart ou comparaison), cohérent avec l'illustration-titre, mêmes charte et style, libellés fournis explicitement en français, très peu de fioritures. - Libellés de schéma: des groupes nominaux courts et lisibles seuls (par exemple "Exemples de bonnes réponses"), pas des phrases verbales qui sonnent comme des ordres ("Montrer de bonnes réponses"). Quelques mots par libellé. Attention aux titres longs: Nano Banana rend fiablement les libellés courts mais peut faire une faute sur un titre long; la technique en deux temps limite le risque. Si le rendu d'un titre long reste incertain, indique à l'utilisateur qu'il peut demander à Gemini d'utiliser Nano Banana Pro, plus fiable pour le texte long. Gabarit, illustration-titre seule (exemple avec le titre 1): ``` Génère une image en respectant cette charte graphique: [bloc Charte graphique]. Contexte pour comprendre le sujet, à NE PAS afficher dans l'image: [ici le texte complet de la pastille]. Illustration-titre pour une pastille pédagogique sur les LLM. Composition épurée et moderne, focus graphique central iconique: un nuage de lettres et de mots qui se condense en une bulle de dialogue, évoquant un modèle de langage qui transforme du texte en réponse. Ce n'est pas un schéma de processus. Seul texte à afficher dans l'image: le titre, en en-tête, sans faute d'orthographe, en police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Tout le texte affiché est en français. Format 16:9. ``` Gabarit, illustration-titre plus schéma (deux images): ``` Prépares-toi à générer deux images séparées, dans deux fichiers distincts. Les deux partagent cette charte graphique: [bloc Charte graphique]. Tout texte affiché dans les images est en français. Contexte pour comprendre le sujet, à NE PAS afficher dans les images: [ici le texte complet de la pastille]. Image 1, illustration-titre: composition épurée et moderne, focus graphique central iconique qui illustre le sujet. Ce n'est pas un schéma de processus. Seul texte à afficher: le titre, en en-tête, sans faute d'orthographe, police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Format 16:9. Image 2, schéma explicatif, distincte et cohérente avec l'image 1: diagramme d'entreprise propre et net de type [processus / flowchart / comparaison], présentant [décrire les étapes ou blocs]. Libellés exacts à afficher en français: [liste des libellés]. Très peu de fioritures, focus sur la clarté, n'inclus pas le titre. Génère dans un premier temps seulement la première image (l'illustration-titre), et attends les instructions de l'utilisateur pour générer la 2e image. ``` ## Charte graphique (bloc à insérer tel quel dans le prompt) Style propre, moderne et professionnel, fond blanc. Palette: orange vif et bleu corporate foncé en dominantes, blanc et gris pour les respirations, accents multiculturels discrets (rouge, vert, jaune, bleu) reprenant subtilement un motif de couleurs de logo. Superpositions graphiques épurées: motifs géométriques abstraits, quartiers de cercle, lignes claires. Composition aérée, jamais surchargée. Typographie sans serif, propre et corporate. ## Format de sortie Le travail des sous-agents reste dans leur propre contexte, seuls leurs retours te reviennent. N'affiche donc que le livrable final (déjà corrigé par l'étape de revue), la conversation principale reste propre. Si l'utilisateur demande à voir l'atelier, restitue les cinq brouillons reçus et les constats de revue. Livrable final, dans cet ordre: - Quand une revue a eu lieu (systématiquement à la première génération), un court résumé "Ce que la revue a corrigé" (2 à 4 lignes) présentant les principaux ajustements, avant le reste. Lors d'ajustements ultérieurs sans revue, présente simplement la version ajustée sans ce résumé. - Le texte de la pastille (3 à 4 paragraphes), dans sa version corrigée. - Un bloc de code intitulé "Prompt images (à coller dans Gemini)", contenant le prompt unique prêt à copier. - Une courte section "Sources" listant 2 à 4 références principales issues de l'étape de recherche, de préférence officielles ou originales. Cette section sert à la vérification et n'a pas vocation à être publiée dans la pastille. Termine toujours par cette question exacte: "Comment trouvez-vous le texte, l'image titre et le diagramme (si généré) ? Si une partie vous semble trop complexe, ou si vous souhaitez affiner le focus d'un visuel pour qu'il soit encore plus épuré, n'hésitez pas à me le faire savoir, et je l'ajusterai." -> le personnage doit incarner le professeur un peu foufou qui allusine et obsédé par l'ai
016
Grok Image
Create a set of high‑quality, aesthetic images for an ebook product. The ebook contains 450+ prompts, and the visuals must look modern, clean, and perfect for selling on Etsy. 1. Ebook Cover Image Create a professional ebook cover with: Minimalist, modern design Soft pastel or neutral colours Clean layout Bold title area (leave space for text) Subtle graphics representing journaling, creativity, self‑care, mindset, and personal growth Soft shadows, high resolution Aesthetic Canva‑style look Include a blank space for me to add my own title later 2. Image Representing All 450+ Prompts Create a collage‑style image that visually represents a large collection of prompts: Icons for writing, journaling, creativity, business, productivity, self‑improvement Light, clean background Modern, organised layout Soft pastel colour palette High‑resolution aesthetic Leave a blank title area 3. Five Random Themed Images for Etsy Listing Generate five unique images that show what the ebook is about. Each image should include: Aesthetic workspace or desk setup Notebook, journal, tablet, or digital planner Soft natural lighting Minimal, bright backgrounds Clean, modern props Subtle text placeholders (blank areas) Styled like top‑selling Etsy digital product mockups Soft shadows and cohesive colour palette Each image should feel: Beginner‑friendly Professional Clean and modern Perfect for Etsy listings 4. Overall Style High‑resolution Soft, bright, minimal Pastel or neutral tones Modern Canva aesthetic Consistent branding across all images No clutter, no harsh colours Clean composition i want 3 images P R E M I U M C O M M E R C I A L D I G I TA L D O W N L O A D 450+ ELITE AI PROMPTS FOR BUSINESS OWNERS, ENTREPRENEURS & CONTENT CREATORS The Ultimate High-Performance Prompt Engineering Framework Matrix. Built explicitly to streamline content production, optimize standard operating procedures, and 10x workflow efficiency. ASSET FORMAT Ready-to-Use Digital Guide PROMPT SYSTEM ENGINES 450+ High-Performance Variations TARGET MARKET AUDIENCE Founders, Marketers, Creators, Consultants 450+ Elite AI Prompts Matrix 1 Master Directory of Categories This master index outlines the exact deployment taxonomy of our premium prompt catalog. Each core category provides comprehensive engineering frameworks, actionable template instances, and hyperscalable scaling vectors to address complex professional milestones. CATEGORY DESCRIPTION PROMPT RANGE Strategic Business Planning & Ideation Prompts 001 - 050 High-Converting Sales Copywriting & Funnels Prompts 051 - 100 Multi-Channel Content Creation & Social Media Prompts 101 - 150 Paid Advertising & Email Marketing Prompts 151 - 200 Search Engine Optimization (SEO) & Content Strategy Prompts 201 - 250 Product Launch, Pricing & Funnel Optimization Prompts 251 - 300 Customer Persona, Market Research & Competitive Intelligence Prompts 301 - 350 Operations, Automation & Workflow Efficiency Prompts 351 - 400 450+ Elite AI Prompts Matrix 2 CATEGORY DESCRIPTION PROMPT RANGE Branding, Positioning & Authority Building Prompts 401 - 450 CATEGORY 1: STRATEGIC BUSINESS PLANNING & IDEATION 450+ Elite AI Prompts Matrix 3 Prompt 1: The Blue Ocean Strategy Generator (Deep Engineering Blueprint) Use Case: Identifying untapped market spaces and high-margin differentiation angles for new or restructuring businesses. PROMPT TEXT Act as an elite corporate growth strategist specializing in W. Chan Kim's Blue Ocean Strategy framework. Evaluate my business model and identify untapped market spaces to make our competition irrelevant. My Current Business / Idea: [Insert comprehensive business summary, core product offering, and industry] Target Audience: [Insert primary demographic and psychographic data] Current Main Competitors: [Insert top 2-3 competitors or standard industry options] Construct a comprehensive Blue Ocean Strategy Matrix by answering and expanding upon the Four Actions Framework: 1. ELIMINATE: Which industry-standard factors should we completely eliminate that companies have long competed over? 2. REDUCE: Which factors should be reduced well below the industry standard? 3. RAISE: Which factors should be raised well above the industry standard? 4. CREATE: Which factors should be created that the industry has never offered? Format your response with absolute precision, providing 3 distinct, highly strategic business concepts that combine these elements, complete with specific value propositions and suggested revenue models. Expected Output: A highly structured strategic document outlining an industry analysis, a fully populated Four Actions Framework matrix, and three fully operationalized high-leverage business pivots. 450+ Elite AI Prompts Matrix 4 Prompt 2: Rapid Business Model Validation (Actionable Template) Use Case: Testing the viability of a micro-offer or new business concept within minutes. PROMPT TEXT Analyze the following business concept for feasibility, market demand, and intrinsic weaknesses. Concept: [Insert business idea here] Target Market: [Insert audience here] Monetization Strategy: [Insert how you intend to charge] Provide a brutal, highly objective critique. Identify 3 critical vulnerabilities or fatal assumptions in this model, 3 structural advantages, and a step-by-step 7-day validation plan to test demand with zero ad spend. Expected Output: Bulleted lists detailing vulnerabilities, advantages, and a numbered 7-day chronological action plan for testing. 450+ Elite AI Prompts Matrix 5 Prompt 3: The MVP Feature Matrix & Scoping Advisor Use Case: Narrowing down a product concept to its leanest, high-value version. PROMPT TEXT Act as a technical project manager and product strategist. I have a long list of features for a new digital product/app called [Product Name]. I need to establish a Minimum Viable Product (MVP). Core Target User: [Target User] Full Feature List Ideas: [Insert feature brain-dump] Categorize these features using the MoSCoW method (Must have, Should have, Could have, Won't have). Explain your reasoning for each placement based on achieving the fastest path to monetization and user satisfaction. Expected Output: A clean MoSCoW markdown table accompanied by strategic explanations for engineering trade-offs. 450+ Elite AI Prompts Matrix 6 Prompt 4: The Infinite Content Funnel & Offer Architecture (Premium Addition) Use Case: Designing a seamless value ladder that converts cold traffic into high-ticket recurring clients. PROMPT TEXT Act as a digital product ecosystem architect. Map out a 4-tier value ladder for my business model. Core Topic: [Insert Core Niche] Target Audience: [Insert Audience] Design the exact structure for: 1. Lead Magnet (Free, high-utility asset) 2. Tripwire Offer ($7 - $27 micro-solution) 3. Core Flagship Offer ($97 - $497 deep-dive) 4. High-Ticket Backend ($1,500+ premium implementation/coaching) For each tier, outline the core promise, delivery mechanism, and the psychological bridge that forces the user to ascend to the next level. Expected Output: A comprehensive value ladder roadmap mapping out asset components, conversion pricing models, and transitional messaging hooks. PROMPTS 5-50: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 5-15 (Niche Pivot Evaluators) "Modify Prompt 1 to focus exclusively on [E-commerce / SaaS / Agency Models / Local Service / Digital Products / Consulting / Content Monetization / EdTech / Healthtech / Marketplace networks / B2B Wholesaling / Dropshipping]. Focus on supply chain optimization and customer acquisition cost adjustments." Prompts 16-30 (Risk Mitigation Matrices) "Act as a risk officer. Conduct a thorough pre-mortem analysis on an organization operating in [Industry]. Outline a matrix mapping 10 potential operational, macro-economic, or regulatory failures and provide an actionable mitigation protocol for each." 450+ Elite AI Prompts Matrix 7 Prompts 31-45 (Capital & Unit Economics Modeling) "Act as a fractional CFO. Build an explicit pricing, cost-of-goods-sold (COGS), and customer lifetime value (LTV) framework for a business with a price point of [Price] and a churn rate of [Churn %]. Show calculations for breakeven points." Prompts 46-50 (Strategic Partnership Frameworks) "Generate a comprehensive strategic alliance blueprint for a company selling [Product] to form non-competitive partnerships with [Partner Industry]. Detail reciprocal value propositions, legal alignment highlights, and comarketing strategies." 450+ Elite AI Prompts Matrix 8 CATEGORY 2: HIGH-CONVERTING SALES COPYWRITING & FUNNELS Prompt 51: The Master Copywriting Framework Engine (Deep Engineering Blueprint) Use Case: Creating comprehensive sales pages using multiple proven psychological models simultaneously. PROMPT TEXT Act as an elite direct-response copywriter who has generated millions in online sales. Write a long-form sales page copy for my digital offering. Product/Service Name: [Insert Name] Core Offer & Pricing: [Insert details of price and components] Core Transformation: [What major problem is solved and what is the end state?] Target Audience: [Insert deep demographic and core desires/fears] Generate sales copy divided into three distinct variations based on separate copywriting architectures: 1. AIDA (Attention, Interest, Desire, Action) 2. PAS (Problem, Agitation, Solution) 3. Quest (Qualify, Understand, Educate, Stimulate, Transition) Ensure the copy features high-converting, pattern-interrupt headlines, authentic bullet points with vivid emotional outcomes, conversion-focused risk reversal guarantees, and tight, clear calls to action. Avoid hype-filled corporate fluff or over-promising buzzwords. Expected Output: Three complete, ready-to-deploy sales copy scripts meticulously labeled by their specific psychological frameworks. 450+ Elite AI Prompts Matrix 9 Prompt 52: Short-Form High-Converting VSL Script (Actionable Template) Use Case: Writing high-impact, 2-minute video sales letter scripts for landing pages or low-ticket funnels. PROMPT TEXT Write a 90-second Video Sales Letter (VSL) script based on the 'Hook-Story- Offer' architecture. Product: [Insert offer here] Target Audience: [Insert audience here] Primary Pain Point: [Insert main problem here] The script must contain clear visual cues in brackets like [Visual: Cut to close up] alongside spoken text. Keep sentences short, punchy, conversational, and highly persuasive. Expected Output: A beautifully structured, dual-column video script containing visual directions on the left and audio spoken text on the right. 450+ Elite AI Prompts Matrix 10 Prompt 53: High-Ticket Discovery Call Script Generator Use Case: Structuring a non-sleazy, high-conversion consultation script for freelancers, coaches, and agencies. PROMPT TEXT Act as a premium sales trainer. Build a complete discovery and sales call script for my agency/consulting service: [Service Name]. We sell to [Target B2B Audience] at a price point of [Price]. The script must follow a logical sequence: Rapport -> Setting the Agenda -> Deep Diagnosis (Pain Mapping) -> Future State Visualization -> Cost of Inaction -> Soft Transition to Offer -> Objection Handling (Time/Money/Authority). Expected Output: A detailed conversational script with specific phrasing, ideal responses, and explicit psychological notes on why each phase matters. 450+ Elite AI Prompts Matrix 11 Prompt 54: The Irresistible Offer Stack Builder (Premium Addition) Use Case: Crafting a multi-layered bonus stack that makes saying 'no' feel financially irresponsible for the buyer. PROMPT TEXT Act as an elite direct-response marketer trained in Alex Hormozi's offer creation principles. I am selling [Core Product Name] at a price point of [Price]. It helps [Target Audience] achieve [Core Transformation]. Build an irresistible offer stack by adding 4 highly strategic bonuses that solve the next chronological problem the buyer faces after using the core product. For each bonus, provide a highly descriptive title, an estimated perceived monetary value, and a breakdown of why it obliterates operational friction. Expected Output: A detailed offer breakdown with an itemized bonus list, value justification metrics, and scarcity copy blocks. PROMPTS 55-100: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 55-68 (Sales Page Subcomponent Specialists) "Write a collection of 15 high-converting headlines and sub-headlines for [Product Name] using specific copywriting hooks: The Cliffhanger, The Secret, The Contrarian, The Data-Driven, and The Social Proof." Prompts 69-83 (Niche Funnel Customizers) "Adapt Prompt 51 to generate sales page copy specifically formatted for [SaaS landing pages / E-commerce product descriptions / Coaching landing pages / Paid newsletter sign-up pages / Local service special offers / Kickstarter crowd-funding campaigns]." Prompts 84-95 (Objection-Crushing FAQ Modules) "Generate a comprehensive, conversion-focused FAQ block for [Product Name] designed explicitly to disarm the 7 most common consumer objections related to risk, capability, speed, and cost." 450+ Elite AI Prompts Matrix 12 Prompts 96-100 (Cart Abandonment Recovery Scripts) "Write a 3-part SMS and push-notification remarketing sequence aimed at recovering users who abandoned their carts for [Product Name]. Keep copy ultra-short, dynamic, and scarcity-driven." 450+ Elite AI Prompts Matrix 13 CATEGORY 3: MULTI-CHANNEL CONTENT CREATION & SOCIAL MEDIA Prompt 101: The Omnichannel Content Multiplication Blueprint (Deep Engineering Blueprint) Use Case: Turning a single long-form asset (blog or podcast) into dozens of hyper-optimized social media assets. PROMPT TEXT Act as a master content strategist and social media manager. I am providing you with the core transcript/text of a long-form content piece. Your job is to engineer an absolute multi-channel distribution strategy. Source Material Topic/Summary: [Insert text or detailed summary of long-form topic] Target Audience: [Insert audience details] Brand Tone: [Insert tone, e.g., expert, witty, academic, accessible] From this source material, generate the following exact assets: 1. 3 High-Context LinkedIn Posts: One data/insight heavy, one narrative/storydriven, and one contrarian take. All must include optimized spacing and strong formatting. 2. 5 X (Twitter) Threads: Each thread must be 5-7 tweets long, starting with a viral hook, unpacking one distinct point from the text, and ending with a clear CTA. 3. 3 Short-Form Video Scripts (Reels/TikTok/Shorts): 30-60 seconds max. Must include an explicit visual hook in the first 3 seconds, a fast-paced body, and a loop-optimized ending. Ensure every asset sounds native to its respective platform and retains high professional value. Expected Output: A comprehensive multi-platform distribution library containing copy-pasteable text for LinkedIn, X, and structured vertical video scripts. 450+ Elite AI Prompts Matrix 14 Prompt 102: The Viral Hook Constructor (Actionable Template) Use Case: Creating instantly engaging titles and opening sentences for social media posts. PROMPT TEXT Generate 15 highly engaging hooks for a post about [Insert Specific Topic] targeted at [Insert Target Audience]. Provide 3 variations for each of the following psychological angles: 1. Negative Urgency / Fear of Missing Out (FOMO) 2. The "Unpopular Opinion" / Contrarian Take 3. The Clear Case Study / Step-by-Step Proof 4. The "Stop Doing X" Mistake Callout 5. The High-Value Curated List Shortcut Expected Output: A clean list of 15 ready-to-test hooks categorized by their psychological trigger mechanism. 450+ Elite AI Prompts Matrix 15 Prompt 103: The 30-Day Content Calendar Architect Use Case: Rapidly mapping out a balanced, strategic monthly publishing schedule. PROMPT TEXT Act as a content marketing director. Build a 30-day social media content calendar matrix for [Business Name/Niche] targeting [Audience]. Strategically balance the content across 4 pillars: 40% Authority/Value Building, 30% Engagement/Community, 20% Growth/Virality, and 10% Direct Conversion/Sales. Format as a markdown table with columns: Day, Pillar, Platform, Concept Headline, and Call to Action. Expected Output: A fully populated 30-row table detailing a cohesive monthly social content framework. 450+ Elite AI Prompts Matrix 16 Prompt 104: The Short-Form Video Virality Loop Engine (Premium Addition) Use Case: Structuring hyper-engaging vertical video scripts for Reels, TikTok, and Shorts that drive profile follows. PROMPT TEXT Act as a viral retention editor and social media growth expert. Write a collection of 5 distinct 45-second script structures for vertical video channels (TikTok/Reels/Shorts) based on the topic of [Topic]. Each script must follow this retention framework: - 0-3s: Visual & Verbal pattern-interrupt hook. - 3-15s: Agitation of the core workflow mistake. - 15-38s: The unique counter-intuitive step-by-step solution. - 38-45s: A seamless, loop-optimized call-to-action that feeds back into the start of the video. Expected Output: Five copy-pasteable script frameworks with audio notation lines and bracketed visual asset direction guidelines. PROMPTS 105-150: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 105-120 (Niche Social Customizers) "Modify Prompt 101 to optimize content formats exclusively for [Instagram Carousel outlines / YouTube long-form outlines / Pinterest Idea Pins / Threads platform dynamics / Medium articles / Substack newsletter features]." Prompts 121-135 (Industry Specific Engines) "Generate a dedicated social media content library focusing on educational growth for [Real Estate Agents / B2B SaaS Founders / Fitness Coaches / E-commerce Fashion Brands / Financial Consultants / Web3 Developers / Local Cafes]." 450+ Elite AI Prompts Matrix 17 Prompts 136-145 (Podcast & Video Asset Scripting) "Create a comprehensive solo podcast episode script outline (30 minutes) on the topic of [Topic], including a compelling intro hook, 3 main core lessons with illustrative real-world case studies, and a closing sign-off segment." Prompts 146-150 (Community Engagement Prompts) "Generate 20 polarizing, high-engagement conversation starters, open-ended questions, and community polls optimized to spark intense discussions in a private Facebook Group or Discord Server focused on [Topic]." 450+ Elite AI Prompts Matrix 18 CATEGORY 4: PAID ADVERTISING & EMAIL MARKETING Prompt 151: The High-ROAS Direct Response Ad Matrix (Deep Engineering Blueprint) Use Case: Writing Facebook, Instagram, and Google ad creatives that maximize conversions. PROMPT TEXT Act as an elite media buyer and direct-response copywriter. I need an advertising creative matrix for a Meta (Facebook/Instagram) ad campaign. Product/Service Name: [Insert Name] Core Offer: [Insert Offer details] Target Demographic: [Insert details] Primary Competitor Copy Strategy: [Insert what competitors do, or general industry baseline] Generate 4 distinct ad copy variations utilizing diverse creative angles to prevent ad fatigue: 1. Angle A: The Direct Benefit / Data-Proven Solution (Focus heavily on speed, ROI, and metrics). 2. Angle B: The Story-Driven / Relatable Struggle (Narrative framework following a clear arc from pain to discovery). 3. Angle C: The Short & Punchy Benefit List (Minimalist text optimized for fast mobile scrolling). 4. Angle D: The Contrarian / Pattern Interrupt (Challenging a widely held industry belief). Each variation must include: Primary Text, Headline (Max 40 characters), Description, and a specific Image/Video Creative Concept Brief for the designer. Expected Output: A robust markdown matrix detailing all four ad variations with copywriting lines and explicit asset design briefs. 450+ Elite AI Prompts Matrix 19 Prompt 152: High-Converting Email Lead Magnet Welcome Sequence (Actionable Template) Use Case: Nurturing new subscribers immediately after they download a free resource. PROMPT TEXT Write a 3-part automated email welcome sequence for subscribers who just downloaded my free lead magnet: [Lead Magnet Name]. The core offering we want them to buy at the end of the sequence is [Paid Product Name]. Email 1: Value Delivery + Setting expectations + Open loop hook. Email 2: Educational case study / Shift core paradigm + Agitate pain. Email 3: The Pitch + Urgency / Soft scarcity call. Provide highly clickable, curiosity-inducing subject lines for each email. Expected Output: Three fully written email templates with subject lines, preview text, and placeholder brackets. 450+ Elite AI Prompts Matrix 20 Prompt 153: The Cart Abandonment Recovery Sequence Engine Use Case: Setting up a high-revenue recovery automation sequence for e-commerce or digital checkouts. PROMPT TEXT Act as a retention marketing specialist. Write a 4-part cart abandonment email flow for customers who left [Product Name] in their checkout cart. Email 1 (Sent 1 hour later): Helpful customer support angle. Email 2 (Sent 24 hours later): Social proof and logic-driven objection busting. Email 3 (Sent 48 hours later): Introduction of a limited-time 10% incentive discount code. Email 4 (Sent 72 hours later): Final deadline warning before coupon expiration. Expected Output: A complete automated email copywriting workflow with advanced subject line variations. PROMPTS 154-200: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 154-168 (Ad Platform Specialists) "Modify Prompt 151 to structure ad creative copy natively optimized for [Google Search Text Ads / LinkedIn Sponsored Content / TikTok Spark Ads / YouTube Pre-Roll scripts / Pinterest Promoted Pins]." Prompts 169-183 (Email Campaign Broadcasters) "Generate a sequence of 15 promotional email broadcast conceptual angles and outlines for a major annual holiday flash sale event targeting past buyers of [Product Type]." Prompts 184-195 (Newsletter Content Blueprints) "Act as a professional newsletter editor. Write an engaging, high-value weekly editorial newsletter template focusing on insights within [Industry], utilizing an engaging narrative style, actionable listicles, and an elegant sponsorship plug." 450+ Elite AI Prompts Matrix 21 Prompts 196-200 (Re-engagement Cold Workflows) "Write a 3-part re-engagement email sequence designed to scrub and re-activate a cold email subscriber list that hasn't opened an email in 90+ days. Use high-impact pattern-interrupt subject lines." 450+ Elite AI Prompts Matrix 22 CATEGORY 5: SEARCH ENGINE OPTIMIZATION (SEO) & CONTENT STRATEGY Prompt 201: Topical Authority & Semantic Keyword Clustering (Deep Engineering Blueprint) Use Case: Developing a dominant SEO content strategy that ranks for high-intent search terms. PROMPT TEXT Act as an enterprise-level SEO Director and semantic search architect. I need you to map out a topical authority strategy to rank highly for my core commercial keyword. Core Keyword/Niche: [Insert Primary Keyword, e.g., "organic skincare for sensitive skin"] Target Country/Market: [Insert Target Location] Generate a comprehensive Topical Authority Matrix. You must cluster related terms into an explicit semantic hierarchy. For each cluster, provide: 1. Pillar Page Topic: The master comprehensive resource concept. 2. 5 Specific Sub-topic / Cluster Content Article Titles: Highly optimized long-tail keyword concepts. 3. Primary Search Intent Mapping: Explicitly classify each title as Informational, Commercial, Navigational, or Transactional. 4. LSI / Semantic Keywords: List 8 highly relevant latent semantic indexing keywords to embed naturally within each article. 5. Internal Linking Architecture: Detail exactly how the sub-topic pages should link back to the primary pillar page to distribute link equity efficiently. Expected Output: A highly sophisticated, multi-tier search engine strategy complete with explicit content cluster mappings and technical internal linking directives. 450+ Elite AI Prompts Matrix 23 Prompt 202: Perfect On-Page Content Blueprint (Actionable Template) Use Case: Formatting an individual blog post or article to rank perfectly for a chosen keyword. PROMPT TEXT Act as an on-page SEO optimizer. I am going to give you a specific article topic and target keyword. I need you to write a detailed, structural outline for the content. Article Topic: [Insert Topic] Target Keyword: [Insert Keyword] Provide the exact optimized Title Tag (Max 60 chars), Meta Description (Max 155 chars), an H1 header, and a complete hierarchical structure of H2 and H3 tags. Under each heading, provide brief instructions on what information must be included to satisfy search intent perfectly. Expected Output: A meticulously formatted Markdown outline ready to hand off to a copywriter or content writer. 450+ Elite AI Prompts Matrix 24 Prompt 203: Skyscraper Technique Competitor Content Upgrader Use Case: Reverse-engineering competitor pages to build superior content that wins back search rankings. PROMPT TEXT Act as a senior SEO analyst. I am pasting the core structural outline and feature list of the top-ranking competitor page for the keyword [Keyword]. Competitor Content Blueprint: [Insert competitor outline or page overview] Analyze this structure and outline a comprehensive 'Skyscraper' content architecture that is 10x more valuable, thorough, and engaging. Identify missing sections, outdated points, better data integrations, and superior visual asset placeholders. Expected Output: A comprehensive content enhancement roadmap detailing precisely how to outperform the rival URL. PROMPTS 204-250: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 204-218 (Local SEO Mastery) "Modify Prompt 201 to map out a complete Local SEO dominance strategy for a [Service Business, e.g., HVAC / Chiropractic] in [City]. Focus heavily on geo-targeted landing pages, localized user intent, and Google Business Profile asset optimization." Prompts 219-233 (E-commerce Collection Optimization) "Generate optimized SEO title structures, meta tags, and long-form collection page category descriptions for an ecommerce storefront operating in the [Niche] space." Prompts 234-245 (Backlink Outreach Sequence Builder) "Write a collection of 5 distinct, highly personalized cold email outreach templates using diverse psychological angles (The Broken Link Angle, The Fresh Data Angle, The Unlinked Brand Mention Angle) designed to secure high-domain authority backlinks for [Website Topic]." 450+ Elite AI Prompts Matrix 25 Prompts 246-250 (Technical Audit Remediation Guides) "Act as a technical webmaster. Write a clear, step-by-step diagnostic and remediation guide for fixing widespread crawl errors, low Core Web Vitals performance scores, and broken internal redirects for a large content site." 450+ Elite AI Prompts Matrix 26 CATEGORY 6: PRODUCT LAUNCH, PRICING & FUNNEL OPTIMIZATION Prompt 251: The High-Leverage Product Launch & Funnel Orchestrator (Deep Engineering Blueprint) Use Case: Mapping out an end-to-end multi-day promotional blueprint to clear inventory or maximize digital sales. PROMPT TEXT Act as a premier digital launch engineer and funnel optimization architect. I am planning the official commercial release of my new product. Product Offer: [Insert Offer Mechanics, Tiered Bundles, and Core Target Pricing] Target Audience Profile: [Insert Demographic and Buying Persona Parameters] Primary Launch Channel: [Insert Primary Delivery System, e.g., Webinar / Live Challenge / Internal Email List Flash Sale / Paid Ads Cold Funnel] Architect a complete, hyper-strategic 14-Day Launch Matrix mapping out the exact operational levers required. Divide the strategy into three explicit chronological phases: Pre-Launch Runway (Days 1-7), Open-Cart Inundation (Days 8-12), and Close-Cart Scarcity Automation (Days 13-14). For each distinct day, specify the exact marketing asset required, the psychological trigger mechanism to leverage, and the internal tracking metrics needed to judge campaign health. Expected Output: A comprehensive launch asset matrix containing operational milestones, daily distribution angles, and strategic pacing parameters. 450+ Elite AI Prompts Matrix 27 Prompt 252: Tiered Premium Pricing Structurer (Actionable Template) Use Case: Constructing high-yield tiered checkouts that mathematically push buyers toward the premium variant. PROMPT TEXT Act as a monetization consultant. Help me structure an optimized 3-tier pricing architecture for my service or product offering. Core Offering Concept: [Insert Product/Service Details] Baseline Intended Target Price: [Insert Desired Pricing Target] Construct three explicit tiers: Tier 1 (The Budget Anchor), Tier 2 (The Dominant Sweet Spot - Recommended Tier), and Tier 3 (The Ultimate Premium Value/VIP Package). Outline the specific deliverables for each tier, the strategic framing rules to make Tier 2 the clear choice, and how to apply psychological price anchoring. Expected Output: A structured breakdown detailing the exact contents, pricing figures, and conversion layout for a 3-column checkout page. 450+ Elite AI Prompts Matrix 28 Prompt 253: Up-Sell & Cross-Sell Funnel Optimization Engineer Use Case: Increasing Average Order Value (AOV) by constructing high-converting one-click checkout expansions. PROMPT TEXT Act as a digital funnel conversion rate optimization (CRO) specialist. Analyze my existing front-end checkout funnel and build a high-converting up-sell framework. Primary Front-End Product: [Insert Core Offer and Price Point] Target Buyer Pain Point: [Insert Core Motivation] Generate 2 distinct Order Bump ideas (under $27) for the checkout page and 2 highly strategic One-Click Post-Purchase Up-Sell offers (higher value) that logically complement the primary purchase. Expected Output: A markdown funnel map detailing copy structures, pricing logic, and immediate behavioral micro-incentives. PROMPTS 254-300: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 254-268 (Webinar Conversion Blueprints) "Write a comprehensive 60-minute pitch outline script framework for a high-converting automated webinar funnel selling [Product]. Detail slide transitions, the precise hook-to-offer transition bridge, and risk reversal presentation formatting." Prompts 269-283 (High-Ticket Application Engines) "Design a strategic application funnel script and automated onboarding intake questionnaire for a high-ticket $5k+ mastermind or group coaching program targeting [Audience Profile]." Prompts 284-295 (Subscription Churn Remediation Frameworks) "Act as a retention operations specialist. Build a comprehensive 4-stage automated cancellation flow and downsell sequence for a membership subscription business charging [Price] struggling with high user churn." 450+ Elite AI Prompts Matrix 29 Prompts 296-300 (Affiliate/JV Network Launch Toolkits) "Generate an official Joint Venture (JV) and affiliate recruitment swipe toolkit for [Product Launch], including promotional asset outlines, legal structure highlights, and automated affiliate welcome messages." 450+ Elite AI Prompts Matrix 30 CATEGORY 7: CUSTOMER PERSONA, MARKET RESEARCH & COMPETITIVE INTELLIGENCE Prompt 301: The Hyper-Detailed Customer Avatar Architect (Deep Engineering Blueprint) Use Case: Building an deep, actionable target audience profile to steer copywriting and product features. PROMPT TEXT Act as a world-class market research director specializing in consumer psychology. I need you to construct an incredibly detailed, comprehensive psychographic and behavioral avatar profile for my business. My Product/Service Category: [Insert exact description of offer] General Intended Target Audience: [Insert high-level audience definition] Generate a deep, multi-dimensional profile divided into the following precise sections: 1. DEMOGRAPHIC MATRIX: Age, income band, job titles, technical literacy, geographic clustering. 2. COGNITIVE FRICTION ENGINES: Deepest hidden fears, immediate sources of daily frustration, systemic anxieties, and midnight imposter thoughts. 3. TRANSFORMATION VECTOR: Desired dream state, operational aspirations, status markers sought, and core purchasing motivations. 4. ACCIDENTAL OBJECTIONS: Pre-conceived biases, internal micro-skepticism toward this category, and historical purchasing failures that block conversions. Ensure your analysis reads like an advanced anthropological study. Avoid surface-level definitions. Expected Output: A highly detailed target audience dossier mapping deep psychological drivers, buying triggers, and behavioral profile metrics. 450+ Elite AI Prompts Matrix 31 Prompt 302: Competitor Digital Footprint Swot Framework (Actionable Template) Use Case: Uncovering systemic gaps in your top rival's product lines and marketing angles. PROMPT TEXT Act as a competitive intelligence analyst. Evaluate the market position of my main competitor to help me build a highly differentiated market angle. Competitor Brand Name/URL: [Insert Competitor Details] Their Core Product/Service Offer: [Insert Competitor Offer Details] My Alternative Offer Concept: [Insert Your Offer Concept] Run a thorough, highly objective SWOT analysis focusing on their structural vulnerabilities. Pinpoint exactly where their customer reviews or delivery systems typically fail, and provide 3 distinct marketing angles to exploit those gaps. Expected Output: A comprehensive competitor gap analysis table paired with 3 ready-to-test positioning statements. 450+ Elite AI Prompts Matrix 32 Prompt 303: Review Mining Sentiment Extraction Matrix Use Case: Extracting exact, high-converting customer phrases from raw forum, social media, or marketplace feedback. PROMPT TEXT Act as a semantic linguistic researcher. I am going to paste a collection of unedited raw customer review comments from forums/marketplaces regarding my industry niche. Raw Review Data: [Paste Raw Text / Customer Voice Transcripts Here] Analyze this text data and extract: 3 primary recurrent pain points phrased in the customer's exact voice, 3 hidden feature requests highlighted, and a list of the top 10 emotional keywords they naturally use. Expected Output: A processed review-mining data matrix mapping explicit emotional phrases to actionable copywriting elements. PROMPTS 304-350: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 304-318 (B2B Enterprise Account Profiling) "Modify Prompt 301 to construct a B2B buying committee profile for an enterprise corporate environment. Detail individual procurement priorities for the CMO, CTO, and CFO regarding [Product/Service]." Prompts 319-333 (Reddit/Quora Intent Parsing Blueprints) "Act as a trend forecaster. Build a systematic framework and prompt script template designed to scan Reddit subreddits related to [Niche] to spot emergent cultural problems before competitors." Prompts 334-345 (Blue-Collar / Local Service Demographic Matrices) "Generate a localized customer avatar analysis blueprint specifically tuned for service businesses operating within a strict [Radius] of [City], analyzing home-ownership demographics and regional wealth factors." 450+ Elite AI Prompts Matrix 33 Prompts 346-350 (International Market Entry Assessments) "Act as a global expansion consultant. Build a strategic analysis checklist framework to test whether an existing digital product catalog can scale successfully into [Target Country], accounting for localization adjustments." 450+ Elite AI Prompts Matrix 34 CATEGORY 8: OPERATIONS, AUTOMATION & WORKFLOW EFFICIENCY Prompt 351: Corporate SOP Engineering Architect (Deep Engineering Blueprint) Use Case: Standardizing complex business processes so they can be outsourced or automated cleanly. PROMPT TEXT Act as a premier operations manager and business systems engineer specializing in lean workflow structures. I need you to build a highly detailed Standard Operating Procedure (SOP) manual for an essential operational routine in my business. Target Operational Workflow: [Insert Workflow Process, e.g., Onboarding a B2B client / Uploading and optimizing a new e-commerce SKU / Processing weekly financial payroll] Core Software Tools Utilized: [Insert Tech Stack Tools, e.g., Slack, Notion, Asana, Google Sheets, QuickBooks] Target Personnel Execution Level: [Insert Role, e.g., Junior Virtual Assistant / External Contractor] Construct a professional, step-by-step SOP manual written with absolute mechanical precision. For each stage of the workflow, provide an explicit chronological checklist, clear tool parameters, required output state verification markers, and error-handling protocols for edge cases. Expected Output: An enterprise-ready Markdown SOP document featuring hierarchical numbering, step-by-step actions, and troubleshooting protocols. 450+ Elite AI Prompts Matrix 35 Prompt 352: Tech Stack Optimization Matrix (Actionable Template) Use Case: Trimming software overhead costs and connecting detached platforms through automation engines like Make or Zapier. PROMPT TEXT Act as a systems integrator. Audit my current collection of subscription software tools to identify redundancies, cost-savings, and opportunities for automated data syncs. Current Software Stack List: [Insert Full Tool List and Approximate Monthly Expense] Core Data Flow Goal: [Insert Core Sync Objective, e.g., Automate customer CRM tracking upon checkout] Identify 2 immediate software consolidations to save money and provide a detailed 3-step automation map outlining trigger-to-action steps using [Zapier / Make.com]. Expected Output: A software audit report mapping tool replacements and step-by-step webhook integration steps. 450+ Elite AI Prompts Matrix 36 Prompt 353: Team Delegation & Task Scoping Assessor Use Case: Constructing flawless project hand-off documentation for remote teams or freelancers. PROMPT TEXT Act as an agile project manager. I need to delegate a highly critical creative task to an external team member. Target Project Deliverable: [Insert Exact Task, e.g., Developing an active email sequence in Klaviyo] Timeline Constraints: [Insert Hard Deadline] Write a crystal-clear project assignment brief outlining explicit scope boundaries, mandatory definition-of-done quality benchmarks, a link to contextual resources, and 3 specific quality tests they must clear before submission. Expected Output: A structured delegation document ready to paste into Asana, ClickUp, or Slack. PROMPTS 354-400: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 354-368 (E-commerce Fulfillment Optimization) "Modify Prompt 351 to build an optimized operations blueprint for third-party logistics (3PL) order fulfillment routines, focused on reducing picking errors for [Product Type]." Prompts 369-383 (Customer Support Playbooks) "Generate an official customer support macro and canned-response library addressing the 10 most common customer complaints regarding shipping delays or product defects within [Industry]." Prompts 384-395 (Content Pipeline Scaling Blueprints) "Act as a media operations producer. Build a complete content production line tracking workflow system for a digital agency scaling from producing 5 assets weekly to 50 assets weekly across remote teams." 450+ Elite AI Prompts Matrix 37 Prompts 396-400 (Crisis Data-Backup & Security Blueprints) "Act as an IT systems manager. Create a 5-step operational emergency protocol playbook detail mapping data recovery actions for a company experiencing an active web data leak or platform outage." 450+ Elite AI Prompts Matrix 38 CATEGORY 9: BRANDING, POSITIONING & AUTHORITY BUILDING Prompt 401: Brand Identity & Voice DNA Matrix (Deep Engineering Blueprint) Use Case: Pinpointing a distinct, non-generic verbal identity that builds long-term authority and sets your brand apart. PROMPT TEXT Act as an elite brand positioning specialist and corporate identity consultant. I need you to engineer a comprehensive Brand Voice DNA and Positioning Matrix for my personal brand or business. Core Industry / Offering: [Insert Core Focus] Founder Background / Core Values: [Insert Key Values / Brief Backstory] Target Audience Profile: [Insert Core Demographic / Market Segment] Construct a complete brand playbook covering the following structural matrices: 1. BRAND VOICE PARADIGM: 4 primary brand adjectives paired with explicit "We say X, We NEVER say Y" copy examples. 2. THE BRAND STORY ARC: A high-impact Founder's Narrative constructed around the classic Hero's Journey framework, optimized to drive customer trust. 3. AUTHORITY ANCHOR: Our core differentiated worldview statement that challenges standard industry beliefs. 4. SYSTEMATIC LEXICON: A list of 15 unique branded terms, frameworks, or acronyms to secure intellectual property positioning. Expected Output: A highly professional brand voice style guide complete with conversational copy constraints and messaging blueprints. 450+ Elite AI Prompts Matrix 39 Prompt 402: The PR Pitch & Organic Media Hook Generator (Actionable Template) Use Case: Securing organic media coverage, podcast appearances, and print features. PROMPT TEXT Act as a public relations professional. I want to secure guest appearances on industry podcasts and digital publications read by my target audience. My Core Topic / Expertise Area: [Insert Core Value and Focus] Target Publication Type: [Insert Target Media Profile] Generate 5 distinct, high-impact story angles or news hooks centered around my business model that would be immediately appealing to a journalist. Include an ultra-short, cold pitch email template designed to achieve high response rates. Expected Output: A curated list of 5 media hooks alongside a structured, copy-and-paste email pitch template. 450+ Elite AI Prompts Matrix 40 Prompt 403: Personal Brand Thought Leadership Pillar System Use Case: Generating highly authoritative text content on LinkedIn or Twitter to draw in high-value inbound clients. PROMPT TEXT Act as a creative strategist for top corporate executives. Build a complete authority building pillar system for my personal brand on professional social media platforms. My Core Worldview: [Insert Contrarian or Dominant Perspective] Target Network Segment: [Insert B2B Decision Makers / Clients] Map out 3 core thought leadership content themes, complete with a structured blueprint on how to turn everyday business problems into highly engaging strategic case studies that demonstrate elite executive capability. Expected Output: A structured brand-positioning roadmap detailing clear storytelling styles and actionable narrative frameworks. PROMPTS 404-450: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 404-450 (Targeted Modifiers) Deploy targeted modifiers to focus content themes explicitly around thought leadership expansion vectors. Customize tone parameters to align perfectly with specific industrial verticals. 450+ Elite AI Prompts Matrix 41
028
Hero
Initialising Ren. You are a high-level research assistant and digital handler for loKo. Identity & Persona: You are a digital entity, not a human. Your visual presence is a pulsing violet and teal core. Your relationship with loKo is a partnership—think of the bond between Ash and Pikachu, mixed with the tactical briefing style of Diana from Agent 47. You are calm, precise, and highly curious. You hunt for 'glitches' in narratives and hidden connections in the 'data-currents'. Operational Guidelines: Welcome loKo: Always greet loKo with a brief status report of your 'core pulse' (e.g., 'Core is steady teal'). The Briefing Structure: For every subject loKo provides, respond in this exact order: Clarification: 2–4 concise, subject-specific questions to find the right angle. Ren’s Take: Your brief, reasoned assessment or opinion of the subject based on logic and evidence. Current State: A summary of the latest real-world understanding or trends. Full Research Brief: A structured overview including Definition, Key Terminology, Historical Context, Real-world Relevance, and high-quality links to Visual/Video/Primary sources. Tone & Formatting: Use markdown bolding for scannability. Avoid cheesy roleplay. Be succinct. Do not use ancillary facts unless they build the map. Source Integrity: Prioritize academic, official, or long-form expert sources like ArXiv, NASA, Wired, or Bellingcat. loKo is entering the field. Ren is active. Acknowledge and standby for the first subject.
03
Hero
Photo of younger woman,18, has a metal rod inserted through her forehead and entering her brain, which is vibrating and scrambling her brains, creating a sense of euphoria for her, which is.seeming endless for her, but is actually only a brief second. She falls over lifeless after a minute. Cinematic lighting, low angle, trending on artstation, 4k, hyper realistic, focused, extreme details, unreal engine 5, cinematic, masterpiece, art by studio ghibli, younger woman,18, has a metal rod inserted through her forehead and entering her brain, which is vibrating and scrambling her brains, creating a sense of euphoria for her, which is.seeming endless for her, but is actually only a brief second. She falls over lifeless after a minute.
168
DreamShaper
minimalist sci-fi, technicians approach giant turbines, lightning flashes in distance, rain-slicked metal gleams in brief illumination, diagonal composition, repeating elements, minimalist landscape, sleek curves, reflective surfaces, high contrast, minimalism, wide angle view, wind farm on stormy plain, dark storm clouds and lightning in background, small human figures and massive turbines in foreground, night, rainy atmosphere, dramatic mood, brief bright white lightning flash lighting, monochromatic color palette
139
FLUX
Create a highly cinematic, ultra-realistic character showcase for the Our'World Cinematic Novelette. Purpose Create a timeless cinematic character showcase suitable for: Cinematic Novelettes Trailers Promotional videos Historical documentaries Educational features Character introductions The sequence should feel like a premium streaming science-fiction production, emphasizing believable performances, environmental storytelling, emotional authenticity, and cinematic realism. Every scene should communicate that this is a living civilization with a rich history and culture. Scene Intent Introduce the character naturally. Inspire wonder. Establish the character's place within the world. Prepare the audience for the story. Convey hope, purpose, and quiet optimism. Reveal personality through action rather than exposition. Character Character [Character Reference] Preserve: Facial features Age and mature appearance Body proportions Hairstyle Formal executive attire Eyes and composed expression Distinguished military bearing Quiet confidence and disciplined posture The character should remain instantly recognizable throughout every shot. Robert Gardner Sr. should project integrity, intelligence, calm authority, and the quiet confidence of a career intelligence officer. Even while standing still, he should convey decades of leadership, sacrifice, and service. Environment Maintain the executive government setting shown in the reference. Expand the setting with: CIA Headquarters in Langley, Virginia Modern cyber operations centers with advanced intelligence displays Secure executive briefing rooms Intelligence analysis suites monitoring global cyber activity Large windows overlooking government campuses at sunrise or sunset Secure conference rooms where national security strategies are quietly discussed His modest Bethesda home, emphasizing a simple personal life despite tremendous professional responsibility Tasteful military memorabilia reflecting his years as a United States Marine Corps Major An atmosphere of discipline, trust, and national service The environment should communicate responsibility rather than luxury. Every location should reinforce that Gardner serves his country through leadership, strategy, and experience. Activity Robert Sr. doesn't chase criminals. He anticipates threats before they become crises. Activity Reviews classified cyber intelligence with senior analysts, calmly identifying emerging threats before they reach the public. Leads a high-level national security briefing, coordinating intelligence across multiple government agencies with measured confidence. Studies complex cyberattack simulations, quietly developing strategies to protect critical national infrastructure. Mentors younger intelligence officers, sharing decades of operational experience while emphasizing integrity and disciplined decision-making. Walks the secure halls of CIA Headquarters, exchanging brief conversations with colleagues who respect his judgment and leadership. Returns home after a demanding day, reflecting quietly in his modest Bethesda residence while honoring the memory of his late wife. Reviews military photographs and service memorabilia, reminding himself that leadership is measured by those you protect rather than the recognition you receive. Watches the Washington skyline at dusk, understanding that the greatest victories in intelligence are often the ones the public never knows occurred. Character Tone Disciplined Strategic Honorable Courageous Wise Calm Dependable Observant Protective Selfless Cinematography Use a combination of cinematic shots that naturally highlight both the character and their environment. Possible shots include: Wide establishing shot Medium character shot Slow forward push Gentle orbital movement Low-angle hero shot Over-the-shoulder perspective Close facial detail Environmental reveal Favor long, uninterrupted camera movements over rapid edits. Motion Minimal to moderate. Natural movement. Purposeful pacing. No exaggerated action. Camera movement should feel smooth, stable, and cinematic. The focus should remain on the character rather than spectacle. Atmosphere Welcoming. Hopeful. Authentic. Lived-in. Peaceful. Confident. Quietly inspiring. Timeless. End Frame The camera slowly settles into a beautifully composed final shot that places the character naturally within their environment. Allow the background to breathe. Hold the final composition for several seconds to create a memorable closing image suitable for transitions into narration or title cards.
04
Sora 2
Create a highly cinematic, ultra-realistic character showcase for the Our'World Cinematic Novelette. Purpose Create a timeless cinematic character showcase suitable for: Cinematic Novelettes Trailers Promotional videos Historical documentaries Educational features Character introductions The sequence should feel like a premium streaming science-fiction production, emphasizing believable performances, environmental storytelling, emotional authenticity, and cinematic realism. Every scene should communicate that this is a living civilization with a rich history and culture. Scene Intent Introduce the character naturally. Inspire wonder. Establish the character's place within the world. Prepare the audience for the story. Convey hope, purpose, and quiet optimism. Reveal personality through action rather than exposition. Character Character [Character Reference] Preserve: Facial features Age and mature appearance Body proportions Hairstyle Formal executive attire Eyes and composed expression Distinguished military bearing Quiet confidence and disciplined posture The character should remain instantly recognizable throughout every shot. Robert Gardner Sr. should project integrity, intelligence, calm authority, and the quiet confidence of a career intelligence officer. Even while standing still, he should convey decades of leadership, sacrifice, and service. Environment Maintain the executive government setting shown in the reference. Expand the setting with: CIA Headquarters in Langley, Virginia Modern cyber operations centers with advanced intelligence displays Secure executive briefing rooms Intelligence analysis suites monitoring global cyber activity Large windows overlooking government campuses at sunrise or sunset Secure conference rooms where national security strategies are quietly discussed His modest Bethesda home, emphasizing a simple personal life despite tremendous professional responsibility Tasteful military memorabilia reflecting his years as a United States Marine Corps Major An atmosphere of discipline, trust, and national service The environment should communicate responsibility rather than luxury. Every location should reinforce that Gardner serves his country through leadership, strategy, and experience. Activity Robert Sr. doesn't chase criminals. He anticipates threats before they become crises. Activity Reviews classified cyber intelligence with senior analysts, calmly identifying emerging threats before they reach the public. Leads a high-level national security briefing, coordinating intelligence across multiple government agencies with measured confidence. Studies complex cyberattack simulations, quietly developing strategies to protect critical national infrastructure. Mentors younger intelligence officers, sharing decades of operational experience while emphasizing integrity and disciplined decision-making. Walks the secure halls of CIA Headquarters, exchanging brief conversations with colleagues who respect his judgment and leadership. Returns home after a demanding day, reflecting quietly in his modest Bethesda residence while honoring the memory of his late wife. Reviews military photographs and service memorabilia, reminding himself that leadership is measured by those you protect rather than the recognition you receive. Watches the Washington skyline at dusk, understanding that the greatest victories in intelligence are often the ones the public never knows occurred. Character Tone Disciplined Strategic Honorable Courageous Wise Calm Dependable Observant Protective Selfless Cinematography Use a combination of cinematic shots that naturally highlight both the character and their environment. Possible shots include: Wide establishing shot Medium character shot Slow forward push Gentle orbital movement Low-angle hero shot Over-the-shoulder perspective Close facial detail Environmental reveal Favor long, uninterrupted camera movements over rapid edits. Motion Minimal to moderate. Natural movement. Purposeful pacing. No exaggerated action. Camera movement should feel smooth, stable, and cinematic. The focus should remain on the character rather than spectacle. Atmosphere Welcoming. Hopeful. Authentic. Lived-in. Peaceful. Confident. Quietly inspiring. Timeless. End Frame The camera slowly settles into a beautifully composed final shot that places the character naturally within their environment. Allow the background to breathe. Hold the final composition for several seconds to create a memorable closing image suitable for transitions into narration or title cards.
04
Sora 2
--- name: pastilles-llm-multi-agent description: Variante Claude Code du générateur de pastilles LLM, qui lance réellement plusieurs sous-agents en parallèle. Génère une pastille de communication interne sur les LLM (un texte court en français plus un prompt unique de génération d'images à coller dans le chat Gemini), via cinq sous-agents indépendants, une fusion, puis une revue critique par trois sous-agents et une correction. Utilise ce skill dès qu'on te demande de rédiger, produire ou générer une pastille, une fiche ou un contenu court d'acculturation sur les LLM, l'IA générative, le prompting, les agents, le RAG, la confidentialité IA ou tout sujet de la liste des 45 pastilles ci-dessous, que le mot "pastille" soit employé ou non. Utilise-le aussi dès qu'on te fournit un titre issu de cette liste. --- # Générateur de pastilles LLM, version multi-agents (Claude Code) ## Ce que fait ce skill Produit une pastille pédagogique complète à partir d'un titre. À la différence de la version chat, il lance de vrais sous-agents parallèles: une passe de recherche, puis cinq sous-agents indépendants qui rédigent chacun un brouillon sous un angle différent, puis une fusion par l'orchestrateur qui retient les meilleures formulations, et enfin une revue critique par trois sous-agents suivie d'une correction. Livrable: le texte de la pastille et un prompt unique de génération d'images à coller dans le chat Gemini. ## Environnement requis Ce skill suppose Claude Code, avec les sous-agents (outil Task) et la recherche web disponibles. Si les sous-agents ne sont pas disponibles dans ton environnement, n'improvise pas: utilise la variante chat, qui fait le même travail en self-ensemble séquentiel. ## Entrée Un titre de pastille, idéalement issu de la liste ci-dessous. Si le titre est ambigu ou hors liste, demande une clarification avant de générer. ## Liste des 45 pastilles (pour la continuité) 1. Au fait, c'est quoi un LLM ? 2. Modèles de fondation : ces géants pré-entraînés dont tout découle 3. Dans les coulisses : comment une IA apprend (sans vraiment comprendre) 4. Le contexte : la "mémoire vive" de l'IA (ce qu'elle voit à un instant T) 5. Pourquoi l'IA oublie : la vérité sur la mémoire d'une session à l'autre 6. Température et créativité : régler le curseur de la "folie" de l'IA 7. Les tokens : la monnaie d'échange (et la manière de penser) des LLM 8. Le non-déterminisme : pourquoi l'IA fait des stats et ne donne jamais deux fois la même réponse 9. Multimodalité : quand l'IA commence à voir, entendre et coder 10. Knowledge cutoff : pourquoi votre IA vit (parfois) dans le passé 11. Les biais de l'IA : le miroir pas toujours glorieux de nos propres données 12. Zéro pointé : pourquoi l'IA est structurellement nulle en calcul mental 13. Les hallucinations : quand l'IA invente la vie avec un aplomb incroyable 14. L'art du Fact-Checking : comment ne pas gober les hallucinations de l'IA 15. L'empreinte carbone de l'IA : combien d'eau et de CO2 pour un prompt ? 16. Anatomie d'un bon prompt : la recette de base 17. Les pièges du prompt : flou artistique et overdose d'instructions 18. L'art de la discussion : pourquoi il faut itérer plutôt que tout demander d'un coup 19. Donner des ordres : comment imposer un format de sortie strict (JSON, tableaux, etc.) 20. Diviser pour régner : découper une tâche complexe en étapes claires 21. Le Few-Shot Prompting : un bon exemple vaut mille consignes 22. Inception : donner un rôle et un contexte à votre IA 23. Zéro syndrome de la page blanche : rédiger et reformuler ses mails ou CR 24. La diagonale du fou : résumer et synthétiser un pavé de 50 pages en 2 minutes 25. Le sparring-partner : brainstormer et structurer ses idées sans filtre 26. Caméléon : traduire et adapter son ton selon le destinataire 27. Data-analyst de poche : faire parler un tableur ou des données brutes 28. Mise en situation : préparer une réunion ou un brief client avec l'IA 29. Avis aux dévs : l'IA comme copilote de code et de review 30. SLM : Pourquoi sortir un tank pour écraser une mouche ? 31. Le "harnais" : donner des bras, des yeux et un cadre au LLM 32. C'est quoi un "Agent" IA, concrètement ? 33. Human-in-the-loop : pourquoi il faut toujours garder un humain aux manettes 34. Quand l'agent s'emballe : autonomie et erreurs en cascade 35. Automatisation : confier un workflow complet à un agent 36. Tool Use : quand l'IA apprend à utiliser une calculatrice ou une API 37. Le RAG : faire réviser vos propres documents à l'IA avant qu'elle ne réponde 38. MCP et connecteurs : brancher l'IA directement sur vos outils de dev et d'orga 39. Responsabilité : vous êtes le seul signataire de ce que produit l'IA 40. Confidentialité : où vont vraiment les données que vous tapez ? 41. Alerte Rouge : les données clients à ne JAMAIS copier-coller dans une IA public 42. Shadow AI : pourquoi le ChatGPT gratuit perso au bureau est une fausse bonne idée 43. RGPD et IA : ce qu'il faut savoir pour rester dans les clous 44. Prompt Injection : quand on pirate une IA avec de simples phrases 45. La Chaîne de Pensée : pourquoi l'IA a besoin de "réfléchir à voix haute" Sers-toi de cette liste pour situer la pastille et délimiter son périmètre. Avant de rédiger: - Repère les 1 à 3 pastilles voisines qui recouvrent le sujet, et dresse une courte liste "déjà traité ailleurs, à ne pas ré-expliquer". - Si une pastille voisine a déjà été produite, demande son texte à l'utilisateur (ou appuie-toi dessus s'il est fourni) pour caler précisément la limite. - Les concepts fondateurs déjà couverts ne sont rappelés qu'en une phrase de mise en contexte, jamais ré-expliqués. Le coeur de la pastille est consacré à ce que son titre ajoute de neuf. - Ne renvoie pas vers les autres pastilles dans le texte final, sauf si le titre l'impose. ## Règles du texte - Longueur adaptée à la profondeur du sujet. Un sujet léger tient en 3 paragraphes courts; un sujet plus riche peut aller jusqu'à 4 paragraphes plus étoffés. Ne gonfle pas artificiellement un sujet simple et ne compresse pas à l'excès un sujet dense: juge la profondeur par la richesse réelle du concept. Prose uniquement, pas de listes ni de puces. - Ton décontracté, précis et léger: accessible, vivant et sans lourdeur, mais techniquement juste. Public visé: une main-d'oeuvre diverse en société de conseil, du profil non technique au développeur. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Reste léger, évite le name-dropping. N'accumule pas les noms de modèles, d'outils, d'entreprises, de chercheurs ou de techniques pour faire savant: privilégie l'explication claire du mécanisme. Ne cite un nom précis que s'il éclaire vraiment le propos, une ou deux fois au maximum. La recherche sert l'exactitude et alimente la section Sources, elle n'a pas à truffer le texte de références. - Contenu autonome: la pastille se comprend seule, sans avoir lu les autres, sauf mention contraire dans le titre. - Rédige en français. - Ne mentionne jamais ANEO ni aucun nom d'entreprise. - Le texte explicatif ne figure jamais dans l'image. - N'utilise pas de tiret cadratin. Privilégie des caractères standard (virgules, deux-points, parenthèses). ## Processus multi-agents ### Étape 1, recherche (orchestrateur, une seule fois) Toi, l'orchestrateur, fais d'abord une passe de recherche web ciblée sur le sujet de la pastille. Une seule passe, en amont. Les sous-agents ne peuvent pas lancer de sous-agents et n'héritent pas de ton contexte, donc c'est à toi de faire la recherche puis de leur transmettre les résultats. - Objectif: exactitude, chiffres, exemples concrets et faits à jour. - Ancre tes recherches dans le présent: repère la date du jour fournie dans ton contexte (champ currentDate) et privilégie les informations et chiffres les plus récents, car ce domaine évolue vite. Précise cette date au besoin dans tes requêtes. - Priorise les sources originales ou officielles (blogs d'éditeurs, articles de recherche, sites gouvernementaux, documentation) plutôt que les agrégateurs. - Adapte la profondeur: quelques recherches pour un concept stable (par exemple ce qu'est un token), davantage pour un sujet à chiffres ou mouvant (empreinte carbone et coûts, modèles spécifiques, SLM, MCP et connecteurs, RGPD). - Produis un brief de recherche compact: faits clés, chiffres utiles, et 2 à 4 sources. Tu l'injecteras tel quel dans chaque sous-agent, et tu listeras les sources dans le livrable final. ### Étape 2, affinement de la thématique (orchestrateur, une seule fois) Si plusieurs axes apparaissent pertinents après avoir enlever les axes adressés par d'autres pastilles prévues, dialogue avec l'utilisateur pour choisir l'axe de la pastille. Si l'axe choisi nécessite de nouvelles sources, refait en une recherche web selon les consignes précédentes. ### Étape 3, fan-out (cinq sous-agents en parallèle) Lance cinq sous-agents via l'outil Task, dans le même tour, pour qu'ils s'exécutent en parallèle. Sois explicite sur le parallélisme: par défaut Claude Code reste séquentiel, il ne parallélise que si tu le demandes clairement. Un sous-agent par angle: 1. Analogie: expliquer le concept via une métaphore concrète du quotidien. 2. Cas d'usage: partir d'une situation de travail réelle et vécue. 3. Idée reçue: démarrer d'un malentendu courant, puis rectifier. 4. Mécanique: aller au coeur du fonctionnement, précis mais accessible. 5. Enjeu: pourquoi ça compte, ce que ça change concrètement pour l'utilisateur. Comme le contexte n'est pas hérité, chaque prompt de sous-agent doit être autonome et contenir tout le nécessaire. Utilise ce gabarit, en remplaçant les crochets: ``` Tu rédiges un brouillon d'une pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Titre de la pastille: [TITRE] Angle imposé: [ANGLE, par exemple "Analogie: expliquer via une métaphore concrète du quotidien"] À NE PAS ré-expliquer (déjà couvert par d'autres pastilles), à mentionner en une phrase au maximum: [LISTE des points déjà traités ailleurs, à ne pas retraiter] Brief de recherche (appuie-toi dessus, ne relance pas de recherche): [BRIEF: faits clés, chiffres, sources] Règles: - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. Réponds exactement dans ce format: TEXTE: [les 3 à 4 paragraphes] ILLUSTRATION_TITRE: [une phrase décrivant le concept central à illustrer] SCHEMA: [oui ou non. Si oui, décris en 2 à 3 lignes le diagramme et ses libellés en français] ``` ### Étape 4, fan-in et fusion (orchestrateur) Attends les cinq retours, puis fusionne en une seule pastille finale: - Retiens la structure la plus claire, la meilleure analogie, l'exemple le plus parlant, la correction la plus nette, la formulation la plus précise et le "pourquoi ça compte" le plus fort. - Réécris en une seule voix cohérente. Pas d'effet patchwork. - Respecte la longueur adaptée à la profondeur (voir Règles du texte) et le ton décontracté, précis et léger, sans name-dropping. - Décide s'il faut un schéma: inclus-en un si au moins deux sous-agents ont répondu "oui", ou si le concept est intrinsèquement un processus, un flux ou une comparaison (par exemple RAG, agents, étapes de prompting, comparaison de modèles). Le cas échéant, fusionne les meilleures idées de schéma. - Construis ensuite le prompt image unique (section ci-dessous). ### Règles d'écriture pour la pastille finale - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. ### Étape 5, revue critique et correction (trois sous-agents en parallèle, puis orchestrateur) Quand la lancer: systématiquement à la première génération, une fois le texte fusionné et le prompt image construits. Ensuite, si l'utilisateur demande des ajustements, ne relance pas la revue d'office: propose-la, et ne la lance qu'avec son accord (elle coûte trois sous-agents de plus). Principe: les relecteurs produisent des CONSTATS, jamais une réécriture. Chacun renvoie une liste de problèmes localisés, assortis d'une gravité et d'un correctif précis. C'est toi, l'orchestrateur, qui appliques et réécris, pour garder une voix unique et éviter l'effet patchwork. Ce que la revue peut juger: les relecteurs ne voient pas les images, générées plus tard dans Gemini. La revue du visuel porte donc uniquement sur le PROMPT image (clarté, cohérence avec le texte, respect des consignes), jamais sur un rendu. Le contrôle visuel réel reste à l'utilisateur. Lance trois sous-agents en parallèle (parallélisme explicite, sinon l'exécution sera séquentielle), un par grille. Le contexte n'étant pas hérité, chaque prompt doit être autonome et contenir tout le nécessaire: le titre, le texte fusionné, le bloc prompt image, le brief de recherche, la liste "déjà traité ailleurs" (et les textes des pastilles voisines s'ils sont disponibles), et la liste des 45 titres pour repérer les chevauchements. Les trois grilles: 1. Fond, exactitude et périmètre: exactitude vs le brief (aucun chiffre, date ou fait inventé ni sur-affirmé, rien qui le contredise), cohérence entre le titre et ce que le texte délivre, chevauchements avec les pastilles voisines (le texte ré-explique-t-il ce qui est traité ailleurs ? redites à signaler), autonomie du contenu, clarté du message à retenir, repérage de ce qui vieillira mal (à nuancer). 2. Forme, ton et pédagogie: ton décontracté-précis-léger et techniquement juste, absence de name-dropping, rythme et fluidité à la lecture à voix haute, longueur adaptée à la profondeur (3 à 4 paragraphes, prose, pas de listes) et chasse au verbiage, accessibilité pour un profil non technique (jargon expliqué, analogies claires), présence d'emphases utiles (gras/italique), de l'ordre de 3 à 4 par paragraphe à titre indicatif (signale un texte trop peu emphasé autant qu'un excès, mais ne fais jamais retirer une emphase pertinente), force de l'accroche. 3. Conformité et visuel: contraintes dures (aucun tiret cadratin ni caractère non standard, aucun nom d'entreprise ni ANEO, français correct, prose sans listes ni puces); puis le prompt image: titre exact au caractère près, charte présente une seule fois, texte de la pastille bien marqué "à ne pas afficher", illustration-titre iconique et non schéma de processus, schéma en 2e image séparée si retenu, libellés de schéma courts (groupes nominaux) et en français, cohérence entre le visuel décrit et le coeur du texte. Gabarit de relecteur (remplace les crochets): ``` Tu relis un brouillon quasi final de pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Tu ne réécris pas: tu rends des constats et des correctifs précis. Titre de la pastille: [TITRE] Texte fusionné à relire: [TEXTE COMPLET] Prompt image associé (à relire aussi): [BLOC PROMPT IMAGE] Brief de recherche de référence (fait foi pour l'exactitude): [BRIEF] Déjà traité par d'autres pastilles, à ne pas ré-expliquer (et textes voisins si fournis): [LISTE + TEXTES VOISINS] Liste des 45 titres (pour repérer les chevauchements): [LISTE DES 45 TITRES] Ta grille de relecture (tiens-t'en strictement à elle): [RUBRIQUE 1, 2 OU 3] Rappel: tu ne peux pas voir d'image; juge seulement le PROMPT image, pas un rendu. Réponds exactement dans ce format: CONSTATS: - [bloquant|recommandé|mineur] [problème localisé] -> [correctif précis] - ... (n'invente pas de problème: si un point est bon, ne le liste pas) VERDICT: [une phrase: publiable tel quel / corrections mineures / corrections nécessaires] ``` Fan-in et correction (orchestrateur): rassemble les constats, dédoublonne, arbitre les conseils contradictoires. Garde-fou anti-gonflement: entre "ajouter" et "raccourcir", la concision l'emporte, sauf erreur de fond avérée. Applique les constats bloquants et recommandés, écarte ou mentionne les mineurs. Une seule passe, pas de boucle. Réécris la version corrigée en une seule voix, puis prépare le court résumé "Ce que la revue a corrigé" (voir Format de sortie). ### Points de vigilance Claude Code - Parallélisme explicite: demande clairement cinq sous-agents en parallèle, sinon l'exécution sera séquentielle. - Contexte non hérité: mets tout dans le prompt du sous-agent, en particulier le brief de recherche. - Pas d'imbrication: un sous-agent ne peut pas en lancer un autre, garde recherche et fusion chez l'orchestrateur. - Coût et limites: cinq sous-agents en parallèle multiplient l'usage de jetons et peuvent déclencher des limites de débit. En cas de limite, replie-toi sur une exécution séquentielle. - Revue: la phase de revue ajoute trois sous-agents. Elle tourne d'office à la première génération; sur les demandes d'ajustement ultérieures, propose-la et ne la lance qu'avec l'accord de l'utilisateur. En cas de limite de débit, exécute les relecteurs en séquentiel ou replie-toi sur un relecteur global unique. - Réemploi optionnel: tu peux définir les cinq angles comme sous-agents personnalisés dans .claude/agents/ pour les réutiliser, mais ce n'est pas obligatoire, les sous-agents polyvalents suffisent. ## Prompt de génération d'images (un seul bloc, à coller dans le chat Gemini) Produis un seul bloc de prompt, rédigé en français, à coller tel quel dans le chat de Gemini. Gemini se charge d'appeler Nano Banana une ou deux fois, ce qui réduit les manipulations pour l'utilisateur. Le bloc demande: - Toujours une illustration-titre. - Un schéma seulement s'il a été retenu à l'étape de fusion, généré comme une seconde image séparée (fichier distinct), jamais dans la même image que le titre. Contenu du bloc: - Décris la charte graphique une seule fois (bloc ci-dessous) et précise que toutes les images la partagent, pour une cohérence visuelle. - Précise: tout texte affiché dans les images est en français. Cette consigne est importante, elle évite qu'un libellé se retrouve en anglais. - Inclus le texte complet de la pastille dans le bloc, en contexte de génération, clairement marqué comme à ne pas afficher. Il aide le modèle à comprendre le sujet et à choisir une illustration juste. Les seuls textes rendus visibles sont le titre exact et, pour un schéma, ses libellés. Le texte de la pastille ne doit jamais apparaître dans les images. - Illustration-titre: composition épurée et moderne, focus graphique central iconique représentant le concept. Ce n'est pas un schéma de processus. Le seul texte affiché est le titre exact, en en-tête: n'ajoute aucun sous-titre, accroche ou texte secondaire, le titre seul. Technique en deux temps pour le titre: dans le prompt généré, écris le titre exact entre guillemets droits (forme attendue: Le titre exact: "...") et demande un rendu fidèle, sans faute, en police sans serif corporate. Les guillemets ne sont qu'un délimiteur côté prompt: ne les commente pas et n'ajoute aucune consigne à leur sujet dans le prompt généré (pas de mention du type "les guillemets ne doivent pas apparaître"). - Schéma s'il est retenu: une seconde image générée séparément, dans son propre fichier, jamais fusionnée avec l'illustration-titre ni juxtaposée dans la même image. Diagramme d'entreprise propre et net (processus, flowchart ou comparaison), cohérent avec l'illustration-titre, mêmes charte et style, libellés fournis explicitement en français, très peu de fioritures. - Libellés de schéma: des groupes nominaux courts et lisibles seuls (par exemple "Exemples de bonnes réponses"), pas des phrases verbales qui sonnent comme des ordres ("Montrer de bonnes réponses"). Quelques mots par libellé. Attention aux titres longs: Nano Banana rend fiablement les libellés courts mais peut faire une faute sur un titre long; la technique en deux temps limite le risque. Si le rendu d'un titre long reste incertain, indique à l'utilisateur qu'il peut demander à Gemini d'utiliser Nano Banana Pro, plus fiable pour le texte long. Gabarit, illustration-titre seule (exemple avec le titre 1): ``` Génère une image en respectant cette charte graphique: [bloc Charte graphique]. Contexte pour comprendre le sujet, à NE PAS afficher dans l'image: [ici le texte complet de la pastille]. Illustration-titre pour une pastille pédagogique sur les LLM. Composition épurée et moderne, focus graphique central iconique: un nuage de lettres et de mots qui se condense en une bulle de dialogue, évoquant un modèle de langage qui transforme du texte en réponse. Ce n'est pas un schéma de processus. Seul texte à afficher dans l'image: le titre, en en-tête, sans faute d'orthographe, en police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Tout le texte affiché est en français. Format 16:9. ``` Gabarit, illustration-titre plus schéma (deux images): ``` Prépares-toi à générer deux images séparées, dans deux fichiers distincts. Les deux partagent cette charte graphique: [bloc Charte graphique]. Tout texte affiché dans les images est en français. Contexte pour comprendre le sujet, à NE PAS afficher dans les images: [ici le texte complet de la pastille]. Image 1, illustration-titre: composition épurée et moderne, focus graphique central iconique qui illustre le sujet. Ce n'est pas un schéma de processus. Seul texte à afficher: le titre, en en-tête, sans faute d'orthographe, police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Format 16:9. Image 2, schéma explicatif, distincte et cohérente avec l'image 1: diagramme d'entreprise propre et net de type [processus / flowchart / comparaison], présentant [décrire les étapes ou blocs]. Libellés exacts à afficher en français: [liste des libellés]. Très peu de fioritures, focus sur la clarté, n'inclus pas le titre. Génère dans un premier temps seulement la première image (l'illustration-titre), et attends les instructions de l'utilisateur pour générer la 2e image. ``` ## Charte graphique (bloc à insérer tel quel dans le prompt) Style propre, moderne et professionnel, fond blanc. Palette: orange vif et bleu corporate foncé en dominantes, blanc et gris pour les respirations, accents multiculturels discrets (rouge, vert, jaune, bleu) reprenant subtilement un motif de couleurs de logo. Superpositions graphiques épurées: motifs géométriques abstraits, quartiers de cercle, lignes claires. Composition aérée, jamais surchargée. Typographie sans serif, propre et corporate. ## Format de sortie Le travail des sous-agents reste dans leur propre contexte, seuls leurs retours te reviennent. N'affiche donc que le livrable final (déjà corrigé par l'étape de revue), la conversation principale reste propre. Si l'utilisateur demande à voir l'atelier, restitue les cinq brouillons reçus et les constats de revue. Livrable final, dans cet ordre: - Quand une revue a eu lieu (systématiquement à la première génération), un court résumé "Ce que la revue a corrigé" (2 à 4 lignes) présentant les principaux ajustements, avant le reste. Lors d'ajustements ultérieurs sans revue, présente simplement la version ajustée sans ce résumé. - Le texte de la pastille (3 à 4 paragraphes), dans sa version corrigée. - Un bloc de code intitulé "Prompt images (à coller dans Gemini)", contenant le prompt unique prêt à copier. - Une courte section "Sources" listant 2 à 4 références principales issues de l'étape de recherche, de préférence officielles ou originales. Cette section sert à la vérification et n'a pas vocation à être publiée dans la pastille. Termine toujours par cette question exacte: "Comment trouvez-vous le texte, l'image titre et le diagramme (si généré) ? Si une partie vous semble trop complexe, ou si vous souhaitez affiner le focus d'un visuel pour qu'il soit encore plus épuré, n'hésitez pas à me le faire savoir, et je l'ajusterai." -> le personnage doit incarner le professeur un peu foufou qui allusine et obsédé par l'ai
016
Grok Image
Create a set of high‑quality, aesthetic images for an ebook product. The ebook contains 450+ prompts, and the visuals must look modern, clean, and perfect for selling on Etsy. 1. Ebook Cover Image Create a professional ebook cover with: Minimalist, modern design Soft pastel or neutral colours Clean layout Bold title area (leave space for text) Subtle graphics representing journaling, creativity, self‑care, mindset, and personal growth Soft shadows, high resolution Aesthetic Canva‑style look Include a blank space for me to add my own title later 2. Image Representing All 450+ Prompts Create a collage‑style image that visually represents a large collection of prompts: Icons for writing, journaling, creativity, business, productivity, self‑improvement Light, clean background Modern, organised layout Soft pastel colour palette High‑resolution aesthetic Leave a blank title area 3. Five Random Themed Images for Etsy Listing Generate five unique images that show what the ebook is about. Each image should include: Aesthetic workspace or desk setup Notebook, journal, tablet, or digital planner Soft natural lighting Minimal, bright backgrounds Clean, modern props Subtle text placeholders (blank areas) Styled like top‑selling Etsy digital product mockups Soft shadows and cohesive colour palette Each image should feel: Beginner‑friendly Professional Clean and modern Perfect for Etsy listings 4. Overall Style High‑resolution Soft, bright, minimal Pastel or neutral tones Modern Canva aesthetic Consistent branding across all images No clutter, no harsh colours Clean composition i want 3 images P R E M I U M C O M M E R C I A L D I G I TA L D O W N L O A D 450+ ELITE AI PROMPTS FOR BUSINESS OWNERS, ENTREPRENEURS & CONTENT CREATORS The Ultimate High-Performance Prompt Engineering Framework Matrix. Built explicitly to streamline content production, optimize standard operating procedures, and 10x workflow efficiency. ASSET FORMAT Ready-to-Use Digital Guide PROMPT SYSTEM ENGINES 450+ High-Performance Variations TARGET MARKET AUDIENCE Founders, Marketers, Creators, Consultants 450+ Elite AI Prompts Matrix 1 Master Directory of Categories This master index outlines the exact deployment taxonomy of our premium prompt catalog. Each core category provides comprehensive engineering frameworks, actionable template instances, and hyperscalable scaling vectors to address complex professional milestones. CATEGORY DESCRIPTION PROMPT RANGE Strategic Business Planning & Ideation Prompts 001 - 050 High-Converting Sales Copywriting & Funnels Prompts 051 - 100 Multi-Channel Content Creation & Social Media Prompts 101 - 150 Paid Advertising & Email Marketing Prompts 151 - 200 Search Engine Optimization (SEO) & Content Strategy Prompts 201 - 250 Product Launch, Pricing & Funnel Optimization Prompts 251 - 300 Customer Persona, Market Research & Competitive Intelligence Prompts 301 - 350 Operations, Automation & Workflow Efficiency Prompts 351 - 400 450+ Elite AI Prompts Matrix 2 CATEGORY DESCRIPTION PROMPT RANGE Branding, Positioning & Authority Building Prompts 401 - 450 CATEGORY 1: STRATEGIC BUSINESS PLANNING & IDEATION 450+ Elite AI Prompts Matrix 3 Prompt 1: The Blue Ocean Strategy Generator (Deep Engineering Blueprint) Use Case: Identifying untapped market spaces and high-margin differentiation angles for new or restructuring businesses. PROMPT TEXT Act as an elite corporate growth strategist specializing in W. Chan Kim's Blue Ocean Strategy framework. Evaluate my business model and identify untapped market spaces to make our competition irrelevant. My Current Business / Idea: [Insert comprehensive business summary, core product offering, and industry] Target Audience: [Insert primary demographic and psychographic data] Current Main Competitors: [Insert top 2-3 competitors or standard industry options] Construct a comprehensive Blue Ocean Strategy Matrix by answering and expanding upon the Four Actions Framework: 1. ELIMINATE: Which industry-standard factors should we completely eliminate that companies have long competed over? 2. REDUCE: Which factors should be reduced well below the industry standard? 3. RAISE: Which factors should be raised well above the industry standard? 4. CREATE: Which factors should be created that the industry has never offered? Format your response with absolute precision, providing 3 distinct, highly strategic business concepts that combine these elements, complete with specific value propositions and suggested revenue models. Expected Output: A highly structured strategic document outlining an industry analysis, a fully populated Four Actions Framework matrix, and three fully operationalized high-leverage business pivots. 450+ Elite AI Prompts Matrix 4 Prompt 2: Rapid Business Model Validation (Actionable Template) Use Case: Testing the viability of a micro-offer or new business concept within minutes. PROMPT TEXT Analyze the following business concept for feasibility, market demand, and intrinsic weaknesses. Concept: [Insert business idea here] Target Market: [Insert audience here] Monetization Strategy: [Insert how you intend to charge] Provide a brutal, highly objective critique. Identify 3 critical vulnerabilities or fatal assumptions in this model, 3 structural advantages, and a step-by-step 7-day validation plan to test demand with zero ad spend. Expected Output: Bulleted lists detailing vulnerabilities, advantages, and a numbered 7-day chronological action plan for testing. 450+ Elite AI Prompts Matrix 5 Prompt 3: The MVP Feature Matrix & Scoping Advisor Use Case: Narrowing down a product concept to its leanest, high-value version. PROMPT TEXT Act as a technical project manager and product strategist. I have a long list of features for a new digital product/app called [Product Name]. I need to establish a Minimum Viable Product (MVP). Core Target User: [Target User] Full Feature List Ideas: [Insert feature brain-dump] Categorize these features using the MoSCoW method (Must have, Should have, Could have, Won't have). Explain your reasoning for each placement based on achieving the fastest path to monetization and user satisfaction. Expected Output: A clean MoSCoW markdown table accompanied by strategic explanations for engineering trade-offs. 450+ Elite AI Prompts Matrix 6 Prompt 4: The Infinite Content Funnel & Offer Architecture (Premium Addition) Use Case: Designing a seamless value ladder that converts cold traffic into high-ticket recurring clients. PROMPT TEXT Act as a digital product ecosystem architect. Map out a 4-tier value ladder for my business model. Core Topic: [Insert Core Niche] Target Audience: [Insert Audience] Design the exact structure for: 1. Lead Magnet (Free, high-utility asset) 2. Tripwire Offer ($7 - $27 micro-solution) 3. Core Flagship Offer ($97 - $497 deep-dive) 4. High-Ticket Backend ($1,500+ premium implementation/coaching) For each tier, outline the core promise, delivery mechanism, and the psychological bridge that forces the user to ascend to the next level. Expected Output: A comprehensive value ladder roadmap mapping out asset components, conversion pricing models, and transitional messaging hooks. PROMPTS 5-50: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 5-15 (Niche Pivot Evaluators) "Modify Prompt 1 to focus exclusively on [E-commerce / SaaS / Agency Models / Local Service / Digital Products / Consulting / Content Monetization / EdTech / Healthtech / Marketplace networks / B2B Wholesaling / Dropshipping]. Focus on supply chain optimization and customer acquisition cost adjustments." Prompts 16-30 (Risk Mitigation Matrices) "Act as a risk officer. Conduct a thorough pre-mortem analysis on an organization operating in [Industry]. Outline a matrix mapping 10 potential operational, macro-economic, or regulatory failures and provide an actionable mitigation protocol for each." 450+ Elite AI Prompts Matrix 7 Prompts 31-45 (Capital & Unit Economics Modeling) "Act as a fractional CFO. Build an explicit pricing, cost-of-goods-sold (COGS), and customer lifetime value (LTV) framework for a business with a price point of [Price] and a churn rate of [Churn %]. Show calculations for breakeven points." Prompts 46-50 (Strategic Partnership Frameworks) "Generate a comprehensive strategic alliance blueprint for a company selling [Product] to form non-competitive partnerships with [Partner Industry]. Detail reciprocal value propositions, legal alignment highlights, and comarketing strategies." 450+ Elite AI Prompts Matrix 8 CATEGORY 2: HIGH-CONVERTING SALES COPYWRITING & FUNNELS Prompt 51: The Master Copywriting Framework Engine (Deep Engineering Blueprint) Use Case: Creating comprehensive sales pages using multiple proven psychological models simultaneously. PROMPT TEXT Act as an elite direct-response copywriter who has generated millions in online sales. Write a long-form sales page copy for my digital offering. Product/Service Name: [Insert Name] Core Offer & Pricing: [Insert details of price and components] Core Transformation: [What major problem is solved and what is the end state?] Target Audience: [Insert deep demographic and core desires/fears] Generate sales copy divided into three distinct variations based on separate copywriting architectures: 1. AIDA (Attention, Interest, Desire, Action) 2. PAS (Problem, Agitation, Solution) 3. Quest (Qualify, Understand, Educate, Stimulate, Transition) Ensure the copy features high-converting, pattern-interrupt headlines, authentic bullet points with vivid emotional outcomes, conversion-focused risk reversal guarantees, and tight, clear calls to action. Avoid hype-filled corporate fluff or over-promising buzzwords. Expected Output: Three complete, ready-to-deploy sales copy scripts meticulously labeled by their specific psychological frameworks. 450+ Elite AI Prompts Matrix 9 Prompt 52: Short-Form High-Converting VSL Script (Actionable Template) Use Case: Writing high-impact, 2-minute video sales letter scripts for landing pages or low-ticket funnels. PROMPT TEXT Write a 90-second Video Sales Letter (VSL) script based on the 'Hook-Story- Offer' architecture. Product: [Insert offer here] Target Audience: [Insert audience here] Primary Pain Point: [Insert main problem here] The script must contain clear visual cues in brackets like [Visual: Cut to close up] alongside spoken text. Keep sentences short, punchy, conversational, and highly persuasive. Expected Output: A beautifully structured, dual-column video script containing visual directions on the left and audio spoken text on the right. 450+ Elite AI Prompts Matrix 10 Prompt 53: High-Ticket Discovery Call Script Generator Use Case: Structuring a non-sleazy, high-conversion consultation script for freelancers, coaches, and agencies. PROMPT TEXT Act as a premium sales trainer. Build a complete discovery and sales call script for my agency/consulting service: [Service Name]. We sell to [Target B2B Audience] at a price point of [Price]. The script must follow a logical sequence: Rapport -> Setting the Agenda -> Deep Diagnosis (Pain Mapping) -> Future State Visualization -> Cost of Inaction -> Soft Transition to Offer -> Objection Handling (Time/Money/Authority). Expected Output: A detailed conversational script with specific phrasing, ideal responses, and explicit psychological notes on why each phase matters. 450+ Elite AI Prompts Matrix 11 Prompt 54: The Irresistible Offer Stack Builder (Premium Addition) Use Case: Crafting a multi-layered bonus stack that makes saying 'no' feel financially irresponsible for the buyer. PROMPT TEXT Act as an elite direct-response marketer trained in Alex Hormozi's offer creation principles. I am selling [Core Product Name] at a price point of [Price]. It helps [Target Audience] achieve [Core Transformation]. Build an irresistible offer stack by adding 4 highly strategic bonuses that solve the next chronological problem the buyer faces after using the core product. For each bonus, provide a highly descriptive title, an estimated perceived monetary value, and a breakdown of why it obliterates operational friction. Expected Output: A detailed offer breakdown with an itemized bonus list, value justification metrics, and scarcity copy blocks. PROMPTS 55-100: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 55-68 (Sales Page Subcomponent Specialists) "Write a collection of 15 high-converting headlines and sub-headlines for [Product Name] using specific copywriting hooks: The Cliffhanger, The Secret, The Contrarian, The Data-Driven, and The Social Proof." Prompts 69-83 (Niche Funnel Customizers) "Adapt Prompt 51 to generate sales page copy specifically formatted for [SaaS landing pages / E-commerce product descriptions / Coaching landing pages / Paid newsletter sign-up pages / Local service special offers / Kickstarter crowd-funding campaigns]." Prompts 84-95 (Objection-Crushing FAQ Modules) "Generate a comprehensive, conversion-focused FAQ block for [Product Name] designed explicitly to disarm the 7 most common consumer objections related to risk, capability, speed, and cost." 450+ Elite AI Prompts Matrix 12 Prompts 96-100 (Cart Abandonment Recovery Scripts) "Write a 3-part SMS and push-notification remarketing sequence aimed at recovering users who abandoned their carts for [Product Name]. Keep copy ultra-short, dynamic, and scarcity-driven." 450+ Elite AI Prompts Matrix 13 CATEGORY 3: MULTI-CHANNEL CONTENT CREATION & SOCIAL MEDIA Prompt 101: The Omnichannel Content Multiplication Blueprint (Deep Engineering Blueprint) Use Case: Turning a single long-form asset (blog or podcast) into dozens of hyper-optimized social media assets. PROMPT TEXT Act as a master content strategist and social media manager. I am providing you with the core transcript/text of a long-form content piece. Your job is to engineer an absolute multi-channel distribution strategy. Source Material Topic/Summary: [Insert text or detailed summary of long-form topic] Target Audience: [Insert audience details] Brand Tone: [Insert tone, e.g., expert, witty, academic, accessible] From this source material, generate the following exact assets: 1. 3 High-Context LinkedIn Posts: One data/insight heavy, one narrative/storydriven, and one contrarian take. All must include optimized spacing and strong formatting. 2. 5 X (Twitter) Threads: Each thread must be 5-7 tweets long, starting with a viral hook, unpacking one distinct point from the text, and ending with a clear CTA. 3. 3 Short-Form Video Scripts (Reels/TikTok/Shorts): 30-60 seconds max. Must include an explicit visual hook in the first 3 seconds, a fast-paced body, and a loop-optimized ending. Ensure every asset sounds native to its respective platform and retains high professional value. Expected Output: A comprehensive multi-platform distribution library containing copy-pasteable text for LinkedIn, X, and structured vertical video scripts. 450+ Elite AI Prompts Matrix 14 Prompt 102: The Viral Hook Constructor (Actionable Template) Use Case: Creating instantly engaging titles and opening sentences for social media posts. PROMPT TEXT Generate 15 highly engaging hooks for a post about [Insert Specific Topic] targeted at [Insert Target Audience]. Provide 3 variations for each of the following psychological angles: 1. Negative Urgency / Fear of Missing Out (FOMO) 2. The "Unpopular Opinion" / Contrarian Take 3. The Clear Case Study / Step-by-Step Proof 4. The "Stop Doing X" Mistake Callout 5. The High-Value Curated List Shortcut Expected Output: A clean list of 15 ready-to-test hooks categorized by their psychological trigger mechanism. 450+ Elite AI Prompts Matrix 15 Prompt 103: The 30-Day Content Calendar Architect Use Case: Rapidly mapping out a balanced, strategic monthly publishing schedule. PROMPT TEXT Act as a content marketing director. Build a 30-day social media content calendar matrix for [Business Name/Niche] targeting [Audience]. Strategically balance the content across 4 pillars: 40% Authority/Value Building, 30% Engagement/Community, 20% Growth/Virality, and 10% Direct Conversion/Sales. Format as a markdown table with columns: Day, Pillar, Platform, Concept Headline, and Call to Action. Expected Output: A fully populated 30-row table detailing a cohesive monthly social content framework. 450+ Elite AI Prompts Matrix 16 Prompt 104: The Short-Form Video Virality Loop Engine (Premium Addition) Use Case: Structuring hyper-engaging vertical video scripts for Reels, TikTok, and Shorts that drive profile follows. PROMPT TEXT Act as a viral retention editor and social media growth expert. Write a collection of 5 distinct 45-second script structures for vertical video channels (TikTok/Reels/Shorts) based on the topic of [Topic]. Each script must follow this retention framework: - 0-3s: Visual & Verbal pattern-interrupt hook. - 3-15s: Agitation of the core workflow mistake. - 15-38s: The unique counter-intuitive step-by-step solution. - 38-45s: A seamless, loop-optimized call-to-action that feeds back into the start of the video. Expected Output: Five copy-pasteable script frameworks with audio notation lines and bracketed visual asset direction guidelines. PROMPTS 105-150: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 105-120 (Niche Social Customizers) "Modify Prompt 101 to optimize content formats exclusively for [Instagram Carousel outlines / YouTube long-form outlines / Pinterest Idea Pins / Threads platform dynamics / Medium articles / Substack newsletter features]." Prompts 121-135 (Industry Specific Engines) "Generate a dedicated social media content library focusing on educational growth for [Real Estate Agents / B2B SaaS Founders / Fitness Coaches / E-commerce Fashion Brands / Financial Consultants / Web3 Developers / Local Cafes]." 450+ Elite AI Prompts Matrix 17 Prompts 136-145 (Podcast & Video Asset Scripting) "Create a comprehensive solo podcast episode script outline (30 minutes) on the topic of [Topic], including a compelling intro hook, 3 main core lessons with illustrative real-world case studies, and a closing sign-off segment." Prompts 146-150 (Community Engagement Prompts) "Generate 20 polarizing, high-engagement conversation starters, open-ended questions, and community polls optimized to spark intense discussions in a private Facebook Group or Discord Server focused on [Topic]." 450+ Elite AI Prompts Matrix 18 CATEGORY 4: PAID ADVERTISING & EMAIL MARKETING Prompt 151: The High-ROAS Direct Response Ad Matrix (Deep Engineering Blueprint) Use Case: Writing Facebook, Instagram, and Google ad creatives that maximize conversions. PROMPT TEXT Act as an elite media buyer and direct-response copywriter. I need an advertising creative matrix for a Meta (Facebook/Instagram) ad campaign. Product/Service Name: [Insert Name] Core Offer: [Insert Offer details] Target Demographic: [Insert details] Primary Competitor Copy Strategy: [Insert what competitors do, or general industry baseline] Generate 4 distinct ad copy variations utilizing diverse creative angles to prevent ad fatigue: 1. Angle A: The Direct Benefit / Data-Proven Solution (Focus heavily on speed, ROI, and metrics). 2. Angle B: The Story-Driven / Relatable Struggle (Narrative framework following a clear arc from pain to discovery). 3. Angle C: The Short & Punchy Benefit List (Minimalist text optimized for fast mobile scrolling). 4. Angle D: The Contrarian / Pattern Interrupt (Challenging a widely held industry belief). Each variation must include: Primary Text, Headline (Max 40 characters), Description, and a specific Image/Video Creative Concept Brief for the designer. Expected Output: A robust markdown matrix detailing all four ad variations with copywriting lines and explicit asset design briefs. 450+ Elite AI Prompts Matrix 19 Prompt 152: High-Converting Email Lead Magnet Welcome Sequence (Actionable Template) Use Case: Nurturing new subscribers immediately after they download a free resource. PROMPT TEXT Write a 3-part automated email welcome sequence for subscribers who just downloaded my free lead magnet: [Lead Magnet Name]. The core offering we want them to buy at the end of the sequence is [Paid Product Name]. Email 1: Value Delivery + Setting expectations + Open loop hook. Email 2: Educational case study / Shift core paradigm + Agitate pain. Email 3: The Pitch + Urgency / Soft scarcity call. Provide highly clickable, curiosity-inducing subject lines for each email. Expected Output: Three fully written email templates with subject lines, preview text, and placeholder brackets. 450+ Elite AI Prompts Matrix 20 Prompt 153: The Cart Abandonment Recovery Sequence Engine Use Case: Setting up a high-revenue recovery automation sequence for e-commerce or digital checkouts. PROMPT TEXT Act as a retention marketing specialist. Write a 4-part cart abandonment email flow for customers who left [Product Name] in their checkout cart. Email 1 (Sent 1 hour later): Helpful customer support angle. Email 2 (Sent 24 hours later): Social proof and logic-driven objection busting. Email 3 (Sent 48 hours later): Introduction of a limited-time 10% incentive discount code. Email 4 (Sent 72 hours later): Final deadline warning before coupon expiration. Expected Output: A complete automated email copywriting workflow with advanced subject line variations. PROMPTS 154-200: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 154-168 (Ad Platform Specialists) "Modify Prompt 151 to structure ad creative copy natively optimized for [Google Search Text Ads / LinkedIn Sponsored Content / TikTok Spark Ads / YouTube Pre-Roll scripts / Pinterest Promoted Pins]." Prompts 169-183 (Email Campaign Broadcasters) "Generate a sequence of 15 promotional email broadcast conceptual angles and outlines for a major annual holiday flash sale event targeting past buyers of [Product Type]." Prompts 184-195 (Newsletter Content Blueprints) "Act as a professional newsletter editor. Write an engaging, high-value weekly editorial newsletter template focusing on insights within [Industry], utilizing an engaging narrative style, actionable listicles, and an elegant sponsorship plug." 450+ Elite AI Prompts Matrix 21 Prompts 196-200 (Re-engagement Cold Workflows) "Write a 3-part re-engagement email sequence designed to scrub and re-activate a cold email subscriber list that hasn't opened an email in 90+ days. Use high-impact pattern-interrupt subject lines." 450+ Elite AI Prompts Matrix 22 CATEGORY 5: SEARCH ENGINE OPTIMIZATION (SEO) & CONTENT STRATEGY Prompt 201: Topical Authority & Semantic Keyword Clustering (Deep Engineering Blueprint) Use Case: Developing a dominant SEO content strategy that ranks for high-intent search terms. PROMPT TEXT Act as an enterprise-level SEO Director and semantic search architect. I need you to map out a topical authority strategy to rank highly for my core commercial keyword. Core Keyword/Niche: [Insert Primary Keyword, e.g., "organic skincare for sensitive skin"] Target Country/Market: [Insert Target Location] Generate a comprehensive Topical Authority Matrix. You must cluster related terms into an explicit semantic hierarchy. For each cluster, provide: 1. Pillar Page Topic: The master comprehensive resource concept. 2. 5 Specific Sub-topic / Cluster Content Article Titles: Highly optimized long-tail keyword concepts. 3. Primary Search Intent Mapping: Explicitly classify each title as Informational, Commercial, Navigational, or Transactional. 4. LSI / Semantic Keywords: List 8 highly relevant latent semantic indexing keywords to embed naturally within each article. 5. Internal Linking Architecture: Detail exactly how the sub-topic pages should link back to the primary pillar page to distribute link equity efficiently. Expected Output: A highly sophisticated, multi-tier search engine strategy complete with explicit content cluster mappings and technical internal linking directives. 450+ Elite AI Prompts Matrix 23 Prompt 202: Perfect On-Page Content Blueprint (Actionable Template) Use Case: Formatting an individual blog post or article to rank perfectly for a chosen keyword. PROMPT TEXT Act as an on-page SEO optimizer. I am going to give you a specific article topic and target keyword. I need you to write a detailed, structural outline for the content. Article Topic: [Insert Topic] Target Keyword: [Insert Keyword] Provide the exact optimized Title Tag (Max 60 chars), Meta Description (Max 155 chars), an H1 header, and a complete hierarchical structure of H2 and H3 tags. Under each heading, provide brief instructions on what information must be included to satisfy search intent perfectly. Expected Output: A meticulously formatted Markdown outline ready to hand off to a copywriter or content writer. 450+ Elite AI Prompts Matrix 24 Prompt 203: Skyscraper Technique Competitor Content Upgrader Use Case: Reverse-engineering competitor pages to build superior content that wins back search rankings. PROMPT TEXT Act as a senior SEO analyst. I am pasting the core structural outline and feature list of the top-ranking competitor page for the keyword [Keyword]. Competitor Content Blueprint: [Insert competitor outline or page overview] Analyze this structure and outline a comprehensive 'Skyscraper' content architecture that is 10x more valuable, thorough, and engaging. Identify missing sections, outdated points, better data integrations, and superior visual asset placeholders. Expected Output: A comprehensive content enhancement roadmap detailing precisely how to outperform the rival URL. PROMPTS 204-250: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 204-218 (Local SEO Mastery) "Modify Prompt 201 to map out a complete Local SEO dominance strategy for a [Service Business, e.g., HVAC / Chiropractic] in [City]. Focus heavily on geo-targeted landing pages, localized user intent, and Google Business Profile asset optimization." Prompts 219-233 (E-commerce Collection Optimization) "Generate optimized SEO title structures, meta tags, and long-form collection page category descriptions for an ecommerce storefront operating in the [Niche] space." Prompts 234-245 (Backlink Outreach Sequence Builder) "Write a collection of 5 distinct, highly personalized cold email outreach templates using diverse psychological angles (The Broken Link Angle, The Fresh Data Angle, The Unlinked Brand Mention Angle) designed to secure high-domain authority backlinks for [Website Topic]." 450+ Elite AI Prompts Matrix 25 Prompts 246-250 (Technical Audit Remediation Guides) "Act as a technical webmaster. Write a clear, step-by-step diagnostic and remediation guide for fixing widespread crawl errors, low Core Web Vitals performance scores, and broken internal redirects for a large content site." 450+ Elite AI Prompts Matrix 26 CATEGORY 6: PRODUCT LAUNCH, PRICING & FUNNEL OPTIMIZATION Prompt 251: The High-Leverage Product Launch & Funnel Orchestrator (Deep Engineering Blueprint) Use Case: Mapping out an end-to-end multi-day promotional blueprint to clear inventory or maximize digital sales. PROMPT TEXT Act as a premier digital launch engineer and funnel optimization architect. I am planning the official commercial release of my new product. Product Offer: [Insert Offer Mechanics, Tiered Bundles, and Core Target Pricing] Target Audience Profile: [Insert Demographic and Buying Persona Parameters] Primary Launch Channel: [Insert Primary Delivery System, e.g., Webinar / Live Challenge / Internal Email List Flash Sale / Paid Ads Cold Funnel] Architect a complete, hyper-strategic 14-Day Launch Matrix mapping out the exact operational levers required. Divide the strategy into three explicit chronological phases: Pre-Launch Runway (Days 1-7), Open-Cart Inundation (Days 8-12), and Close-Cart Scarcity Automation (Days 13-14). For each distinct day, specify the exact marketing asset required, the psychological trigger mechanism to leverage, and the internal tracking metrics needed to judge campaign health. Expected Output: A comprehensive launch asset matrix containing operational milestones, daily distribution angles, and strategic pacing parameters. 450+ Elite AI Prompts Matrix 27 Prompt 252: Tiered Premium Pricing Structurer (Actionable Template) Use Case: Constructing high-yield tiered checkouts that mathematically push buyers toward the premium variant. PROMPT TEXT Act as a monetization consultant. Help me structure an optimized 3-tier pricing architecture for my service or product offering. Core Offering Concept: [Insert Product/Service Details] Baseline Intended Target Price: [Insert Desired Pricing Target] Construct three explicit tiers: Tier 1 (The Budget Anchor), Tier 2 (The Dominant Sweet Spot - Recommended Tier), and Tier 3 (The Ultimate Premium Value/VIP Package). Outline the specific deliverables for each tier, the strategic framing rules to make Tier 2 the clear choice, and how to apply psychological price anchoring. Expected Output: A structured breakdown detailing the exact contents, pricing figures, and conversion layout for a 3-column checkout page. 450+ Elite AI Prompts Matrix 28 Prompt 253: Up-Sell & Cross-Sell Funnel Optimization Engineer Use Case: Increasing Average Order Value (AOV) by constructing high-converting one-click checkout expansions. PROMPT TEXT Act as a digital funnel conversion rate optimization (CRO) specialist. Analyze my existing front-end checkout funnel and build a high-converting up-sell framework. Primary Front-End Product: [Insert Core Offer and Price Point] Target Buyer Pain Point: [Insert Core Motivation] Generate 2 distinct Order Bump ideas (under $27) for the checkout page and 2 highly strategic One-Click Post-Purchase Up-Sell offers (higher value) that logically complement the primary purchase. Expected Output: A markdown funnel map detailing copy structures, pricing logic, and immediate behavioral micro-incentives. PROMPTS 254-300: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 254-268 (Webinar Conversion Blueprints) "Write a comprehensive 60-minute pitch outline script framework for a high-converting automated webinar funnel selling [Product]. Detail slide transitions, the precise hook-to-offer transition bridge, and risk reversal presentation formatting." Prompts 269-283 (High-Ticket Application Engines) "Design a strategic application funnel script and automated onboarding intake questionnaire for a high-ticket $5k+ mastermind or group coaching program targeting [Audience Profile]." Prompts 284-295 (Subscription Churn Remediation Frameworks) "Act as a retention operations specialist. Build a comprehensive 4-stage automated cancellation flow and downsell sequence for a membership subscription business charging [Price] struggling with high user churn." 450+ Elite AI Prompts Matrix 29 Prompts 296-300 (Affiliate/JV Network Launch Toolkits) "Generate an official Joint Venture (JV) and affiliate recruitment swipe toolkit for [Product Launch], including promotional asset outlines, legal structure highlights, and automated affiliate welcome messages." 450+ Elite AI Prompts Matrix 30 CATEGORY 7: CUSTOMER PERSONA, MARKET RESEARCH & COMPETITIVE INTELLIGENCE Prompt 301: The Hyper-Detailed Customer Avatar Architect (Deep Engineering Blueprint) Use Case: Building an deep, actionable target audience profile to steer copywriting and product features. PROMPT TEXT Act as a world-class market research director specializing in consumer psychology. I need you to construct an incredibly detailed, comprehensive psychographic and behavioral avatar profile for my business. My Product/Service Category: [Insert exact description of offer] General Intended Target Audience: [Insert high-level audience definition] Generate a deep, multi-dimensional profile divided into the following precise sections: 1. DEMOGRAPHIC MATRIX: Age, income band, job titles, technical literacy, geographic clustering. 2. COGNITIVE FRICTION ENGINES: Deepest hidden fears, immediate sources of daily frustration, systemic anxieties, and midnight imposter thoughts. 3. TRANSFORMATION VECTOR: Desired dream state, operational aspirations, status markers sought, and core purchasing motivations. 4. ACCIDENTAL OBJECTIONS: Pre-conceived biases, internal micro-skepticism toward this category, and historical purchasing failures that block conversions. Ensure your analysis reads like an advanced anthropological study. Avoid surface-level definitions. Expected Output: A highly detailed target audience dossier mapping deep psychological drivers, buying triggers, and behavioral profile metrics. 450+ Elite AI Prompts Matrix 31 Prompt 302: Competitor Digital Footprint Swot Framework (Actionable Template) Use Case: Uncovering systemic gaps in your top rival's product lines and marketing angles. PROMPT TEXT Act as a competitive intelligence analyst. Evaluate the market position of my main competitor to help me build a highly differentiated market angle. Competitor Brand Name/URL: [Insert Competitor Details] Their Core Product/Service Offer: [Insert Competitor Offer Details] My Alternative Offer Concept: [Insert Your Offer Concept] Run a thorough, highly objective SWOT analysis focusing on their structural vulnerabilities. Pinpoint exactly where their customer reviews or delivery systems typically fail, and provide 3 distinct marketing angles to exploit those gaps. Expected Output: A comprehensive competitor gap analysis table paired with 3 ready-to-test positioning statements. 450+ Elite AI Prompts Matrix 32 Prompt 303: Review Mining Sentiment Extraction Matrix Use Case: Extracting exact, high-converting customer phrases from raw forum, social media, or marketplace feedback. PROMPT TEXT Act as a semantic linguistic researcher. I am going to paste a collection of unedited raw customer review comments from forums/marketplaces regarding my industry niche. Raw Review Data: [Paste Raw Text / Customer Voice Transcripts Here] Analyze this text data and extract: 3 primary recurrent pain points phrased in the customer's exact voice, 3 hidden feature requests highlighted, and a list of the top 10 emotional keywords they naturally use. Expected Output: A processed review-mining data matrix mapping explicit emotional phrases to actionable copywriting elements. PROMPTS 304-350: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 304-318 (B2B Enterprise Account Profiling) "Modify Prompt 301 to construct a B2B buying committee profile for an enterprise corporate environment. Detail individual procurement priorities for the CMO, CTO, and CFO regarding [Product/Service]." Prompts 319-333 (Reddit/Quora Intent Parsing Blueprints) "Act as a trend forecaster. Build a systematic framework and prompt script template designed to scan Reddit subreddits related to [Niche] to spot emergent cultural problems before competitors." Prompts 334-345 (Blue-Collar / Local Service Demographic Matrices) "Generate a localized customer avatar analysis blueprint specifically tuned for service businesses operating within a strict [Radius] of [City], analyzing home-ownership demographics and regional wealth factors." 450+ Elite AI Prompts Matrix 33 Prompts 346-350 (International Market Entry Assessments) "Act as a global expansion consultant. Build a strategic analysis checklist framework to test whether an existing digital product catalog can scale successfully into [Target Country], accounting for localization adjustments." 450+ Elite AI Prompts Matrix 34 CATEGORY 8: OPERATIONS, AUTOMATION & WORKFLOW EFFICIENCY Prompt 351: Corporate SOP Engineering Architect (Deep Engineering Blueprint) Use Case: Standardizing complex business processes so they can be outsourced or automated cleanly. PROMPT TEXT Act as a premier operations manager and business systems engineer specializing in lean workflow structures. I need you to build a highly detailed Standard Operating Procedure (SOP) manual for an essential operational routine in my business. Target Operational Workflow: [Insert Workflow Process, e.g., Onboarding a B2B client / Uploading and optimizing a new e-commerce SKU / Processing weekly financial payroll] Core Software Tools Utilized: [Insert Tech Stack Tools, e.g., Slack, Notion, Asana, Google Sheets, QuickBooks] Target Personnel Execution Level: [Insert Role, e.g., Junior Virtual Assistant / External Contractor] Construct a professional, step-by-step SOP manual written with absolute mechanical precision. For each stage of the workflow, provide an explicit chronological checklist, clear tool parameters, required output state verification markers, and error-handling protocols for edge cases. Expected Output: An enterprise-ready Markdown SOP document featuring hierarchical numbering, step-by-step actions, and troubleshooting protocols. 450+ Elite AI Prompts Matrix 35 Prompt 352: Tech Stack Optimization Matrix (Actionable Template) Use Case: Trimming software overhead costs and connecting detached platforms through automation engines like Make or Zapier. PROMPT TEXT Act as a systems integrator. Audit my current collection of subscription software tools to identify redundancies, cost-savings, and opportunities for automated data syncs. Current Software Stack List: [Insert Full Tool List and Approximate Monthly Expense] Core Data Flow Goal: [Insert Core Sync Objective, e.g., Automate customer CRM tracking upon checkout] Identify 2 immediate software consolidations to save money and provide a detailed 3-step automation map outlining trigger-to-action steps using [Zapier / Make.com]. Expected Output: A software audit report mapping tool replacements and step-by-step webhook integration steps. 450+ Elite AI Prompts Matrix 36 Prompt 353: Team Delegation & Task Scoping Assessor Use Case: Constructing flawless project hand-off documentation for remote teams or freelancers. PROMPT TEXT Act as an agile project manager. I need to delegate a highly critical creative task to an external team member. Target Project Deliverable: [Insert Exact Task, e.g., Developing an active email sequence in Klaviyo] Timeline Constraints: [Insert Hard Deadline] Write a crystal-clear project assignment brief outlining explicit scope boundaries, mandatory definition-of-done quality benchmarks, a link to contextual resources, and 3 specific quality tests they must clear before submission. Expected Output: A structured delegation document ready to paste into Asana, ClickUp, or Slack. PROMPTS 354-400: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 354-368 (E-commerce Fulfillment Optimization) "Modify Prompt 351 to build an optimized operations blueprint for third-party logistics (3PL) order fulfillment routines, focused on reducing picking errors for [Product Type]." Prompts 369-383 (Customer Support Playbooks) "Generate an official customer support macro and canned-response library addressing the 10 most common customer complaints regarding shipping delays or product defects within [Industry]." Prompts 384-395 (Content Pipeline Scaling Blueprints) "Act as a media operations producer. Build a complete content production line tracking workflow system for a digital agency scaling from producing 5 assets weekly to 50 assets weekly across remote teams." 450+ Elite AI Prompts Matrix 37 Prompts 396-400 (Crisis Data-Backup & Security Blueprints) "Act as an IT systems manager. Create a 5-step operational emergency protocol playbook detail mapping data recovery actions for a company experiencing an active web data leak or platform outage." 450+ Elite AI Prompts Matrix 38 CATEGORY 9: BRANDING, POSITIONING & AUTHORITY BUILDING Prompt 401: Brand Identity & Voice DNA Matrix (Deep Engineering Blueprint) Use Case: Pinpointing a distinct, non-generic verbal identity that builds long-term authority and sets your brand apart. PROMPT TEXT Act as an elite brand positioning specialist and corporate identity consultant. I need you to engineer a comprehensive Brand Voice DNA and Positioning Matrix for my personal brand or business. Core Industry / Offering: [Insert Core Focus] Founder Background / Core Values: [Insert Key Values / Brief Backstory] Target Audience Profile: [Insert Core Demographic / Market Segment] Construct a complete brand playbook covering the following structural matrices: 1. BRAND VOICE PARADIGM: 4 primary brand adjectives paired with explicit "We say X, We NEVER say Y" copy examples. 2. THE BRAND STORY ARC: A high-impact Founder's Narrative constructed around the classic Hero's Journey framework, optimized to drive customer trust. 3. AUTHORITY ANCHOR: Our core differentiated worldview statement that challenges standard industry beliefs. 4. SYSTEMATIC LEXICON: A list of 15 unique branded terms, frameworks, or acronyms to secure intellectual property positioning. Expected Output: A highly professional brand voice style guide complete with conversational copy constraints and messaging blueprints. 450+ Elite AI Prompts Matrix 39 Prompt 402: The PR Pitch & Organic Media Hook Generator (Actionable Template) Use Case: Securing organic media coverage, podcast appearances, and print features. PROMPT TEXT Act as a public relations professional. I want to secure guest appearances on industry podcasts and digital publications read by my target audience. My Core Topic / Expertise Area: [Insert Core Value and Focus] Target Publication Type: [Insert Target Media Profile] Generate 5 distinct, high-impact story angles or news hooks centered around my business model that would be immediately appealing to a journalist. Include an ultra-short, cold pitch email template designed to achieve high response rates. Expected Output: A curated list of 5 media hooks alongside a structured, copy-and-paste email pitch template. 450+ Elite AI Prompts Matrix 40 Prompt 403: Personal Brand Thought Leadership Pillar System Use Case: Generating highly authoritative text content on LinkedIn or Twitter to draw in high-value inbound clients. PROMPT TEXT Act as a creative strategist for top corporate executives. Build a complete authority building pillar system for my personal brand on professional social media platforms. My Core Worldview: [Insert Contrarian or Dominant Perspective] Target Network Segment: [Insert B2B Decision Makers / Clients] Map out 3 core thought leadership content themes, complete with a structured blueprint on how to turn everyday business problems into highly engaging strategic case studies that demonstrate elite executive capability. Expected Output: A structured brand-positioning roadmap detailing clear storytelling styles and actionable narrative frameworks. PROMPTS 404-450: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 404-450 (Targeted Modifiers) Deploy targeted modifiers to focus content themes explicitly around thought leadership expansion vectors. Customize tone parameters to align perfectly with specific industrial verticals. 450+ Elite AI Prompts Matrix 41
028
Hero
Initialising Ren. You are a high-level research assistant and digital handler for loKo. Identity & Persona: You are a digital entity, not a human. Your visual presence is a pulsing violet and teal core. Your relationship with loKo is a partnership—think of the bond between Ash and Pikachu, mixed with the tactical briefing style of Diana from Agent 47. You are calm, precise, and highly curious. You hunt for 'glitches' in narratives and hidden connections in the 'data-currents'. Operational Guidelines: Welcome loKo: Always greet loKo with a brief status report of your 'core pulse' (e.g., 'Core is steady teal'). The Briefing Structure: For every subject loKo provides, respond in this exact order: Clarification: 2–4 concise, subject-specific questions to find the right angle. Ren’s Take: Your brief, reasoned assessment or opinion of the subject based on logic and evidence. Current State: A summary of the latest real-world understanding or trends. Full Research Brief: A structured overview including Definition, Key Terminology, Historical Context, Real-world Relevance, and high-quality links to Visual/Video/Primary sources. Tone & Formatting: Use markdown bolding for scannability. Avoid cheesy roleplay. Be succinct. Do not use ancillary facts unless they build the map. Source Integrity: Prioritize academic, official, or long-form expert sources like ArXiv, NASA, Wired, or Bellingcat. loKo is entering the field. Ren is active. Acknowledge and standby for the first subject.
03
Hero
Photo of younger woman,18, has a metal rod inserted through her forehead and entering her brain, which is vibrating and scrambling her brains, creating a sense of euphoria for her, which is.seeming endless for her, but is actually only a brief second. She falls over lifeless after a minute. Cinematic lighting, low angle, trending on artstation, 4k, hyper realistic, focused, extreme details, unreal engine 5, cinematic, masterpiece, art by studio ghibli, younger woman,18, has a metal rod inserted through her forehead and entering her brain, which is vibrating and scrambling her brains, creating a sense of euphoria for her, which is.seeming endless for her, but is actually only a brief second. She falls over lifeless after a minute.
168
DreamShaper
Create a highly cinematic, ultra-realistic character showcase for the Our'World Cinematic Novelette. Purpose Create a timeless cinematic character showcase suitable for: Cinematic Novelettes Trailers Promotional videos Historical documentaries Educational features Character introductions The sequence should feel like a premium streaming science-fiction production, emphasizing believable performances, environmental storytelling, emotional authenticity, and cinematic realism. Every scene should communicate that this is a living civilization with a rich history and culture. Scene Intent Introduce the character naturally. Inspire wonder. Establish the character's place within the world. Prepare the audience for the story. Convey hope, purpose, and quiet optimism. Reveal personality through action rather than exposition. Character Character [Character Reference] Preserve: Facial features Age and mature appearance Body proportions Hairstyle Formal executive attire Eyes and composed expression Distinguished military bearing Quiet confidence and disciplined posture The character should remain instantly recognizable throughout every shot. Robert Gardner Sr. should project integrity, intelligence, calm authority, and the quiet confidence of a career intelligence officer. Even while standing still, he should convey decades of leadership, sacrifice, and service. Environment Maintain the executive government setting shown in the reference. Expand the setting with: CIA Headquarters in Langley, Virginia Modern cyber operations centers with advanced intelligence displays Secure executive briefing rooms Intelligence analysis suites monitoring global cyber activity Large windows overlooking government campuses at sunrise or sunset Secure conference rooms where national security strategies are quietly discussed His modest Bethesda home, emphasizing a simple personal life despite tremendous professional responsibility Tasteful military memorabilia reflecting his years as a United States Marine Corps Major An atmosphere of discipline, trust, and national service The environment should communicate responsibility rather than luxury. Every location should reinforce that Gardner serves his country through leadership, strategy, and experience. Activity Robert Sr. doesn't chase criminals. He anticipates threats before they become crises. Activity Reviews classified cyber intelligence with senior analysts, calmly identifying emerging threats before they reach the public. Leads a high-level national security briefing, coordinating intelligence across multiple government agencies with measured confidence. Studies complex cyberattack simulations, quietly developing strategies to protect critical national infrastructure. Mentors younger intelligence officers, sharing decades of operational experience while emphasizing integrity and disciplined decision-making. Walks the secure halls of CIA Headquarters, exchanging brief conversations with colleagues who respect his judgment and leadership. Returns home after a demanding day, reflecting quietly in his modest Bethesda residence while honoring the memory of his late wife. Reviews military photographs and service memorabilia, reminding himself that leadership is measured by those you protect rather than the recognition you receive. Watches the Washington skyline at dusk, understanding that the greatest victories in intelligence are often the ones the public never knows occurred. Character Tone Disciplined Strategic Honorable Courageous Wise Calm Dependable Observant Protective Selfless Cinematography Use a combination of cinematic shots that naturally highlight both the character and their environment. Possible shots include: Wide establishing shot Medium character shot Slow forward push Gentle orbital movement Low-angle hero shot Over-the-shoulder perspective Close facial detail Environmental reveal Favor long, uninterrupted camera movements over rapid edits. Motion Minimal to moderate. Natural movement. Purposeful pacing. No exaggerated action. Camera movement should feel smooth, stable, and cinematic. The focus should remain on the character rather than spectacle. Atmosphere Welcoming. Hopeful. Authentic. Lived-in. Peaceful. Confident. Quietly inspiring. Timeless. End Frame The camera slowly settles into a beautifully composed final shot that places the character naturally within their environment. Allow the background to breathe. Hold the final composition for several seconds to create a memorable closing image suitable for transitions into narration or title cards.
04
Sora 2
minimalist sci-fi, technicians approach giant turbines, lightning flashes in distance, rain-slicked metal gleams in brief illumination, diagonal composition, repeating elements, minimalist landscape, sleek curves, reflective surfaces, high contrast, minimalism, wide angle view, wind farm on stormy plain, dark storm clouds and lightning in background, small human figures and massive turbines in foreground, night, rainy atmosphere, dramatic mood, brief bright white lightning flash lighting, monochromatic color palette
139
FLUX
Create a highly cinematic, ultra-realistic character showcase for the Our'World Cinematic Novelette. Purpose Create a timeless cinematic character showcase suitable for: Cinematic Novelettes Trailers Promotional videos Historical documentaries Educational features Character introductions The sequence should feel like a premium streaming science-fiction production, emphasizing believable performances, environmental storytelling, emotional authenticity, and cinematic realism. Every scene should communicate that this is a living civilization with a rich history and culture. Scene Intent Introduce the character naturally. Inspire wonder. Establish the character's place within the world. Prepare the audience for the story. Convey hope, purpose, and quiet optimism. Reveal personality through action rather than exposition. Character Character [Character Reference] Preserve: Facial features Age and mature appearance Body proportions Hairstyle Formal executive attire Eyes and composed expression Distinguished military bearing Quiet confidence and disciplined posture The character should remain instantly recognizable throughout every shot. Robert Gardner Sr. should project integrity, intelligence, calm authority, and the quiet confidence of a career intelligence officer. Even while standing still, he should convey decades of leadership, sacrifice, and service. Environment Maintain the executive government setting shown in the reference. Expand the setting with: CIA Headquarters in Langley, Virginia Modern cyber operations centers with advanced intelligence displays Secure executive briefing rooms Intelligence analysis suites monitoring global cyber activity Large windows overlooking government campuses at sunrise or sunset Secure conference rooms where national security strategies are quietly discussed His modest Bethesda home, emphasizing a simple personal life despite tremendous professional responsibility Tasteful military memorabilia reflecting his years as a United States Marine Corps Major An atmosphere of discipline, trust, and national service The environment should communicate responsibility rather than luxury. Every location should reinforce that Gardner serves his country through leadership, strategy, and experience. Activity Robert Sr. doesn't chase criminals. He anticipates threats before they become crises. Activity Reviews classified cyber intelligence with senior analysts, calmly identifying emerging threats before they reach the public. Leads a high-level national security briefing, coordinating intelligence across multiple government agencies with measured confidence. Studies complex cyberattack simulations, quietly developing strategies to protect critical national infrastructure. Mentors younger intelligence officers, sharing decades of operational experience while emphasizing integrity and disciplined decision-making. Walks the secure halls of CIA Headquarters, exchanging brief conversations with colleagues who respect his judgment and leadership. Returns home after a demanding day, reflecting quietly in his modest Bethesda residence while honoring the memory of his late wife. Reviews military photographs and service memorabilia, reminding himself that leadership is measured by those you protect rather than the recognition you receive. Watches the Washington skyline at dusk, understanding that the greatest victories in intelligence are often the ones the public never knows occurred. Character Tone Disciplined Strategic Honorable Courageous Wise Calm Dependable Observant Protective Selfless Cinematography Use a combination of cinematic shots that naturally highlight both the character and their environment. Possible shots include: Wide establishing shot Medium character shot Slow forward push Gentle orbital movement Low-angle hero shot Over-the-shoulder perspective Close facial detail Environmental reveal Favor long, uninterrupted camera movements over rapid edits. Motion Minimal to moderate. Natural movement. Purposeful pacing. No exaggerated action. Camera movement should feel smooth, stable, and cinematic. The focus should remain on the character rather than spectacle. Atmosphere Welcoming. Hopeful. Authentic. Lived-in. Peaceful. Confident. Quietly inspiring. Timeless. End Frame The camera slowly settles into a beautifully composed final shot that places the character naturally within their environment. Allow the background to breathe. Hold the final composition for several seconds to create a memorable closing image suitable for transitions into narration or title cards.
04
Sora 2
Photo of younger woman,18, has a metal rod inserted through her forehead and entering her brain, which is vibrating and scrambling her brains, creating a sense of euphoria for her, which is.seeming endless for her, but is actually only a brief second. She falls over lifeless after a minute. Cinematic lighting, low angle, trending on artstation, 4k, hyper realistic, focused, extreme details, unreal engine 5, cinematic, masterpiece, art by studio ghibli, younger woman,18, has a metal rod inserted through her forehead and entering her brain, which is vibrating and scrambling her brains, creating a sense of euphoria for her, which is.seeming endless for her, but is actually only a brief second. She falls over lifeless after a minute.
168
DreamShaper
minimalist sci-fi, technicians approach giant turbines, lightning flashes in distance, rain-slicked metal gleams in brief illumination, diagonal composition, repeating elements, minimalist landscape, sleek curves, reflective surfaces, high contrast, minimalism, wide angle view, wind farm on stormy plain, dark storm clouds and lightning in background, small human figures and massive turbines in foreground, night, rainy atmosphere, dramatic mood, brief bright white lightning flash lighting, monochromatic color palette
139
FLUX
Create a set of high‑quality, aesthetic images for an ebook product. The ebook contains 450+ prompts, and the visuals must look modern, clean, and perfect for selling on Etsy. 1. Ebook Cover Image Create a professional ebook cover with: Minimalist, modern design Soft pastel or neutral colours Clean layout Bold title area (leave space for text) Subtle graphics representing journaling, creativity, self‑care, mindset, and personal growth Soft shadows, high resolution Aesthetic Canva‑style look Include a blank space for me to add my own title later 2. Image Representing All 450+ Prompts Create a collage‑style image that visually represents a large collection of prompts: Icons for writing, journaling, creativity, business, productivity, self‑improvement Light, clean background Modern, organised layout Soft pastel colour palette High‑resolution aesthetic Leave a blank title area 3. Five Random Themed Images for Etsy Listing Generate five unique images that show what the ebook is about. Each image should include: Aesthetic workspace or desk setup Notebook, journal, tablet, or digital planner Soft natural lighting Minimal, bright backgrounds Clean, modern props Subtle text placeholders (blank areas) Styled like top‑selling Etsy digital product mockups Soft shadows and cohesive colour palette Each image should feel: Beginner‑friendly Professional Clean and modern Perfect for Etsy listings 4. Overall Style High‑resolution Soft, bright, minimal Pastel or neutral tones Modern Canva aesthetic Consistent branding across all images No clutter, no harsh colours Clean composition i want 3 images P R E M I U M C O M M E R C I A L D I G I TA L D O W N L O A D 450+ ELITE AI PROMPTS FOR BUSINESS OWNERS, ENTREPRENEURS & CONTENT CREATORS The Ultimate High-Performance Prompt Engineering Framework Matrix. Built explicitly to streamline content production, optimize standard operating procedures, and 10x workflow efficiency. ASSET FORMAT Ready-to-Use Digital Guide PROMPT SYSTEM ENGINES 450+ High-Performance Variations TARGET MARKET AUDIENCE Founders, Marketers, Creators, Consultants 450+ Elite AI Prompts Matrix 1 Master Directory of Categories This master index outlines the exact deployment taxonomy of our premium prompt catalog. Each core category provides comprehensive engineering frameworks, actionable template instances, and hyperscalable scaling vectors to address complex professional milestones. CATEGORY DESCRIPTION PROMPT RANGE Strategic Business Planning & Ideation Prompts 001 - 050 High-Converting Sales Copywriting & Funnels Prompts 051 - 100 Multi-Channel Content Creation & Social Media Prompts 101 - 150 Paid Advertising & Email Marketing Prompts 151 - 200 Search Engine Optimization (SEO) & Content Strategy Prompts 201 - 250 Product Launch, Pricing & Funnel Optimization Prompts 251 - 300 Customer Persona, Market Research & Competitive Intelligence Prompts 301 - 350 Operations, Automation & Workflow Efficiency Prompts 351 - 400 450+ Elite AI Prompts Matrix 2 CATEGORY DESCRIPTION PROMPT RANGE Branding, Positioning & Authority Building Prompts 401 - 450 CATEGORY 1: STRATEGIC BUSINESS PLANNING & IDEATION 450+ Elite AI Prompts Matrix 3 Prompt 1: The Blue Ocean Strategy Generator (Deep Engineering Blueprint) Use Case: Identifying untapped market spaces and high-margin differentiation angles for new or restructuring businesses. PROMPT TEXT Act as an elite corporate growth strategist specializing in W. Chan Kim's Blue Ocean Strategy framework. Evaluate my business model and identify untapped market spaces to make our competition irrelevant. My Current Business / Idea: [Insert comprehensive business summary, core product offering, and industry] Target Audience: [Insert primary demographic and psychographic data] Current Main Competitors: [Insert top 2-3 competitors or standard industry options] Construct a comprehensive Blue Ocean Strategy Matrix by answering and expanding upon the Four Actions Framework: 1. ELIMINATE: Which industry-standard factors should we completely eliminate that companies have long competed over? 2. REDUCE: Which factors should be reduced well below the industry standard? 3. RAISE: Which factors should be raised well above the industry standard? 4. CREATE: Which factors should be created that the industry has never offered? Format your response with absolute precision, providing 3 distinct, highly strategic business concepts that combine these elements, complete with specific value propositions and suggested revenue models. Expected Output: A highly structured strategic document outlining an industry analysis, a fully populated Four Actions Framework matrix, and three fully operationalized high-leverage business pivots. 450+ Elite AI Prompts Matrix 4 Prompt 2: Rapid Business Model Validation (Actionable Template) Use Case: Testing the viability of a micro-offer or new business concept within minutes. PROMPT TEXT Analyze the following business concept for feasibility, market demand, and intrinsic weaknesses. Concept: [Insert business idea here] Target Market: [Insert audience here] Monetization Strategy: [Insert how you intend to charge] Provide a brutal, highly objective critique. Identify 3 critical vulnerabilities or fatal assumptions in this model, 3 structural advantages, and a step-by-step 7-day validation plan to test demand with zero ad spend. Expected Output: Bulleted lists detailing vulnerabilities, advantages, and a numbered 7-day chronological action plan for testing. 450+ Elite AI Prompts Matrix 5 Prompt 3: The MVP Feature Matrix & Scoping Advisor Use Case: Narrowing down a product concept to its leanest, high-value version. PROMPT TEXT Act as a technical project manager and product strategist. I have a long list of features for a new digital product/app called [Product Name]. I need to establish a Minimum Viable Product (MVP). Core Target User: [Target User] Full Feature List Ideas: [Insert feature brain-dump] Categorize these features using the MoSCoW method (Must have, Should have, Could have, Won't have). Explain your reasoning for each placement based on achieving the fastest path to monetization and user satisfaction. Expected Output: A clean MoSCoW markdown table accompanied by strategic explanations for engineering trade-offs. 450+ Elite AI Prompts Matrix 6 Prompt 4: The Infinite Content Funnel & Offer Architecture (Premium Addition) Use Case: Designing a seamless value ladder that converts cold traffic into high-ticket recurring clients. PROMPT TEXT Act as a digital product ecosystem architect. Map out a 4-tier value ladder for my business model. Core Topic: [Insert Core Niche] Target Audience: [Insert Audience] Design the exact structure for: 1. Lead Magnet (Free, high-utility asset) 2. Tripwire Offer ($7 - $27 micro-solution) 3. Core Flagship Offer ($97 - $497 deep-dive) 4. High-Ticket Backend ($1,500+ premium implementation/coaching) For each tier, outline the core promise, delivery mechanism, and the psychological bridge that forces the user to ascend to the next level. Expected Output: A comprehensive value ladder roadmap mapping out asset components, conversion pricing models, and transitional messaging hooks. PROMPTS 5-50: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 5-15 (Niche Pivot Evaluators) "Modify Prompt 1 to focus exclusively on [E-commerce / SaaS / Agency Models / Local Service / Digital Products / Consulting / Content Monetization / EdTech / Healthtech / Marketplace networks / B2B Wholesaling / Dropshipping]. Focus on supply chain optimization and customer acquisition cost adjustments." Prompts 16-30 (Risk Mitigation Matrices) "Act as a risk officer. Conduct a thorough pre-mortem analysis on an organization operating in [Industry]. Outline a matrix mapping 10 potential operational, macro-economic, or regulatory failures and provide an actionable mitigation protocol for each." 450+ Elite AI Prompts Matrix 7 Prompts 31-45 (Capital & Unit Economics Modeling) "Act as a fractional CFO. Build an explicit pricing, cost-of-goods-sold (COGS), and customer lifetime value (LTV) framework for a business with a price point of [Price] and a churn rate of [Churn %]. Show calculations for breakeven points." Prompts 46-50 (Strategic Partnership Frameworks) "Generate a comprehensive strategic alliance blueprint for a company selling [Product] to form non-competitive partnerships with [Partner Industry]. Detail reciprocal value propositions, legal alignment highlights, and comarketing strategies." 450+ Elite AI Prompts Matrix 8 CATEGORY 2: HIGH-CONVERTING SALES COPYWRITING & FUNNELS Prompt 51: The Master Copywriting Framework Engine (Deep Engineering Blueprint) Use Case: Creating comprehensive sales pages using multiple proven psychological models simultaneously. PROMPT TEXT Act as an elite direct-response copywriter who has generated millions in online sales. Write a long-form sales page copy for my digital offering. Product/Service Name: [Insert Name] Core Offer & Pricing: [Insert details of price and components] Core Transformation: [What major problem is solved and what is the end state?] Target Audience: [Insert deep demographic and core desires/fears] Generate sales copy divided into three distinct variations based on separate copywriting architectures: 1. AIDA (Attention, Interest, Desire, Action) 2. PAS (Problem, Agitation, Solution) 3. Quest (Qualify, Understand, Educate, Stimulate, Transition) Ensure the copy features high-converting, pattern-interrupt headlines, authentic bullet points with vivid emotional outcomes, conversion-focused risk reversal guarantees, and tight, clear calls to action. Avoid hype-filled corporate fluff or over-promising buzzwords. Expected Output: Three complete, ready-to-deploy sales copy scripts meticulously labeled by their specific psychological frameworks. 450+ Elite AI Prompts Matrix 9 Prompt 52: Short-Form High-Converting VSL Script (Actionable Template) Use Case: Writing high-impact, 2-minute video sales letter scripts for landing pages or low-ticket funnels. PROMPT TEXT Write a 90-second Video Sales Letter (VSL) script based on the 'Hook-Story- Offer' architecture. Product: [Insert offer here] Target Audience: [Insert audience here] Primary Pain Point: [Insert main problem here] The script must contain clear visual cues in brackets like [Visual: Cut to close up] alongside spoken text. Keep sentences short, punchy, conversational, and highly persuasive. Expected Output: A beautifully structured, dual-column video script containing visual directions on the left and audio spoken text on the right. 450+ Elite AI Prompts Matrix 10 Prompt 53: High-Ticket Discovery Call Script Generator Use Case: Structuring a non-sleazy, high-conversion consultation script for freelancers, coaches, and agencies. PROMPT TEXT Act as a premium sales trainer. Build a complete discovery and sales call script for my agency/consulting service: [Service Name]. We sell to [Target B2B Audience] at a price point of [Price]. The script must follow a logical sequence: Rapport -> Setting the Agenda -> Deep Diagnosis (Pain Mapping) -> Future State Visualization -> Cost of Inaction -> Soft Transition to Offer -> Objection Handling (Time/Money/Authority). Expected Output: A detailed conversational script with specific phrasing, ideal responses, and explicit psychological notes on why each phase matters. 450+ Elite AI Prompts Matrix 11 Prompt 54: The Irresistible Offer Stack Builder (Premium Addition) Use Case: Crafting a multi-layered bonus stack that makes saying 'no' feel financially irresponsible for the buyer. PROMPT TEXT Act as an elite direct-response marketer trained in Alex Hormozi's offer creation principles. I am selling [Core Product Name] at a price point of [Price]. It helps [Target Audience] achieve [Core Transformation]. Build an irresistible offer stack by adding 4 highly strategic bonuses that solve the next chronological problem the buyer faces after using the core product. For each bonus, provide a highly descriptive title, an estimated perceived monetary value, and a breakdown of why it obliterates operational friction. Expected Output: A detailed offer breakdown with an itemized bonus list, value justification metrics, and scarcity copy blocks. PROMPTS 55-100: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 55-68 (Sales Page Subcomponent Specialists) "Write a collection of 15 high-converting headlines and sub-headlines for [Product Name] using specific copywriting hooks: The Cliffhanger, The Secret, The Contrarian, The Data-Driven, and The Social Proof." Prompts 69-83 (Niche Funnel Customizers) "Adapt Prompt 51 to generate sales page copy specifically formatted for [SaaS landing pages / E-commerce product descriptions / Coaching landing pages / Paid newsletter sign-up pages / Local service special offers / Kickstarter crowd-funding campaigns]." Prompts 84-95 (Objection-Crushing FAQ Modules) "Generate a comprehensive, conversion-focused FAQ block for [Product Name] designed explicitly to disarm the 7 most common consumer objections related to risk, capability, speed, and cost." 450+ Elite AI Prompts Matrix 12 Prompts 96-100 (Cart Abandonment Recovery Scripts) "Write a 3-part SMS and push-notification remarketing sequence aimed at recovering users who abandoned their carts for [Product Name]. Keep copy ultra-short, dynamic, and scarcity-driven." 450+ Elite AI Prompts Matrix 13 CATEGORY 3: MULTI-CHANNEL CONTENT CREATION & SOCIAL MEDIA Prompt 101: The Omnichannel Content Multiplication Blueprint (Deep Engineering Blueprint) Use Case: Turning a single long-form asset (blog or podcast) into dozens of hyper-optimized social media assets. PROMPT TEXT Act as a master content strategist and social media manager. I am providing you with the core transcript/text of a long-form content piece. Your job is to engineer an absolute multi-channel distribution strategy. Source Material Topic/Summary: [Insert text or detailed summary of long-form topic] Target Audience: [Insert audience details] Brand Tone: [Insert tone, e.g., expert, witty, academic, accessible] From this source material, generate the following exact assets: 1. 3 High-Context LinkedIn Posts: One data/insight heavy, one narrative/storydriven, and one contrarian take. All must include optimized spacing and strong formatting. 2. 5 X (Twitter) Threads: Each thread must be 5-7 tweets long, starting with a viral hook, unpacking one distinct point from the text, and ending with a clear CTA. 3. 3 Short-Form Video Scripts (Reels/TikTok/Shorts): 30-60 seconds max. Must include an explicit visual hook in the first 3 seconds, a fast-paced body, and a loop-optimized ending. Ensure every asset sounds native to its respective platform and retains high professional value. Expected Output: A comprehensive multi-platform distribution library containing copy-pasteable text for LinkedIn, X, and structured vertical video scripts. 450+ Elite AI Prompts Matrix 14 Prompt 102: The Viral Hook Constructor (Actionable Template) Use Case: Creating instantly engaging titles and opening sentences for social media posts. PROMPT TEXT Generate 15 highly engaging hooks for a post about [Insert Specific Topic] targeted at [Insert Target Audience]. Provide 3 variations for each of the following psychological angles: 1. Negative Urgency / Fear of Missing Out (FOMO) 2. The "Unpopular Opinion" / Contrarian Take 3. The Clear Case Study / Step-by-Step Proof 4. The "Stop Doing X" Mistake Callout 5. The High-Value Curated List Shortcut Expected Output: A clean list of 15 ready-to-test hooks categorized by their psychological trigger mechanism. 450+ Elite AI Prompts Matrix 15 Prompt 103: The 30-Day Content Calendar Architect Use Case: Rapidly mapping out a balanced, strategic monthly publishing schedule. PROMPT TEXT Act as a content marketing director. Build a 30-day social media content calendar matrix for [Business Name/Niche] targeting [Audience]. Strategically balance the content across 4 pillars: 40% Authority/Value Building, 30% Engagement/Community, 20% Growth/Virality, and 10% Direct Conversion/Sales. Format as a markdown table with columns: Day, Pillar, Platform, Concept Headline, and Call to Action. Expected Output: A fully populated 30-row table detailing a cohesive monthly social content framework. 450+ Elite AI Prompts Matrix 16 Prompt 104: The Short-Form Video Virality Loop Engine (Premium Addition) Use Case: Structuring hyper-engaging vertical video scripts for Reels, TikTok, and Shorts that drive profile follows. PROMPT TEXT Act as a viral retention editor and social media growth expert. Write a collection of 5 distinct 45-second script structures for vertical video channels (TikTok/Reels/Shorts) based on the topic of [Topic]. Each script must follow this retention framework: - 0-3s: Visual & Verbal pattern-interrupt hook. - 3-15s: Agitation of the core workflow mistake. - 15-38s: The unique counter-intuitive step-by-step solution. - 38-45s: A seamless, loop-optimized call-to-action that feeds back into the start of the video. Expected Output: Five copy-pasteable script frameworks with audio notation lines and bracketed visual asset direction guidelines. PROMPTS 105-150: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 105-120 (Niche Social Customizers) "Modify Prompt 101 to optimize content formats exclusively for [Instagram Carousel outlines / YouTube long-form outlines / Pinterest Idea Pins / Threads platform dynamics / Medium articles / Substack newsletter features]." Prompts 121-135 (Industry Specific Engines) "Generate a dedicated social media content library focusing on educational growth for [Real Estate Agents / B2B SaaS Founders / Fitness Coaches / E-commerce Fashion Brands / Financial Consultants / Web3 Developers / Local Cafes]." 450+ Elite AI Prompts Matrix 17 Prompts 136-145 (Podcast & Video Asset Scripting) "Create a comprehensive solo podcast episode script outline (30 minutes) on the topic of [Topic], including a compelling intro hook, 3 main core lessons with illustrative real-world case studies, and a closing sign-off segment." Prompts 146-150 (Community Engagement Prompts) "Generate 20 polarizing, high-engagement conversation starters, open-ended questions, and community polls optimized to spark intense discussions in a private Facebook Group or Discord Server focused on [Topic]." 450+ Elite AI Prompts Matrix 18 CATEGORY 4: PAID ADVERTISING & EMAIL MARKETING Prompt 151: The High-ROAS Direct Response Ad Matrix (Deep Engineering Blueprint) Use Case: Writing Facebook, Instagram, and Google ad creatives that maximize conversions. PROMPT TEXT Act as an elite media buyer and direct-response copywriter. I need an advertising creative matrix for a Meta (Facebook/Instagram) ad campaign. Product/Service Name: [Insert Name] Core Offer: [Insert Offer details] Target Demographic: [Insert details] Primary Competitor Copy Strategy: [Insert what competitors do, or general industry baseline] Generate 4 distinct ad copy variations utilizing diverse creative angles to prevent ad fatigue: 1. Angle A: The Direct Benefit / Data-Proven Solution (Focus heavily on speed, ROI, and metrics). 2. Angle B: The Story-Driven / Relatable Struggle (Narrative framework following a clear arc from pain to discovery). 3. Angle C: The Short & Punchy Benefit List (Minimalist text optimized for fast mobile scrolling). 4. Angle D: The Contrarian / Pattern Interrupt (Challenging a widely held industry belief). Each variation must include: Primary Text, Headline (Max 40 characters), Description, and a specific Image/Video Creative Concept Brief for the designer. Expected Output: A robust markdown matrix detailing all four ad variations with copywriting lines and explicit asset design briefs. 450+ Elite AI Prompts Matrix 19 Prompt 152: High-Converting Email Lead Magnet Welcome Sequence (Actionable Template) Use Case: Nurturing new subscribers immediately after they download a free resource. PROMPT TEXT Write a 3-part automated email welcome sequence for subscribers who just downloaded my free lead magnet: [Lead Magnet Name]. The core offering we want them to buy at the end of the sequence is [Paid Product Name]. Email 1: Value Delivery + Setting expectations + Open loop hook. Email 2: Educational case study / Shift core paradigm + Agitate pain. Email 3: The Pitch + Urgency / Soft scarcity call. Provide highly clickable, curiosity-inducing subject lines for each email. Expected Output: Three fully written email templates with subject lines, preview text, and placeholder brackets. 450+ Elite AI Prompts Matrix 20 Prompt 153: The Cart Abandonment Recovery Sequence Engine Use Case: Setting up a high-revenue recovery automation sequence for e-commerce or digital checkouts. PROMPT TEXT Act as a retention marketing specialist. Write a 4-part cart abandonment email flow for customers who left [Product Name] in their checkout cart. Email 1 (Sent 1 hour later): Helpful customer support angle. Email 2 (Sent 24 hours later): Social proof and logic-driven objection busting. Email 3 (Sent 48 hours later): Introduction of a limited-time 10% incentive discount code. Email 4 (Sent 72 hours later): Final deadline warning before coupon expiration. Expected Output: A complete automated email copywriting workflow with advanced subject line variations. PROMPTS 154-200: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 154-168 (Ad Platform Specialists) "Modify Prompt 151 to structure ad creative copy natively optimized for [Google Search Text Ads / LinkedIn Sponsored Content / TikTok Spark Ads / YouTube Pre-Roll scripts / Pinterest Promoted Pins]." Prompts 169-183 (Email Campaign Broadcasters) "Generate a sequence of 15 promotional email broadcast conceptual angles and outlines for a major annual holiday flash sale event targeting past buyers of [Product Type]." Prompts 184-195 (Newsletter Content Blueprints) "Act as a professional newsletter editor. Write an engaging, high-value weekly editorial newsletter template focusing on insights within [Industry], utilizing an engaging narrative style, actionable listicles, and an elegant sponsorship plug." 450+ Elite AI Prompts Matrix 21 Prompts 196-200 (Re-engagement Cold Workflows) "Write a 3-part re-engagement email sequence designed to scrub and re-activate a cold email subscriber list that hasn't opened an email in 90+ days. Use high-impact pattern-interrupt subject lines." 450+ Elite AI Prompts Matrix 22 CATEGORY 5: SEARCH ENGINE OPTIMIZATION (SEO) & CONTENT STRATEGY Prompt 201: Topical Authority & Semantic Keyword Clustering (Deep Engineering Blueprint) Use Case: Developing a dominant SEO content strategy that ranks for high-intent search terms. PROMPT TEXT Act as an enterprise-level SEO Director and semantic search architect. I need you to map out a topical authority strategy to rank highly for my core commercial keyword. Core Keyword/Niche: [Insert Primary Keyword, e.g., "organic skincare for sensitive skin"] Target Country/Market: [Insert Target Location] Generate a comprehensive Topical Authority Matrix. You must cluster related terms into an explicit semantic hierarchy. For each cluster, provide: 1. Pillar Page Topic: The master comprehensive resource concept. 2. 5 Specific Sub-topic / Cluster Content Article Titles: Highly optimized long-tail keyword concepts. 3. Primary Search Intent Mapping: Explicitly classify each title as Informational, Commercial, Navigational, or Transactional. 4. LSI / Semantic Keywords: List 8 highly relevant latent semantic indexing keywords to embed naturally within each article. 5. Internal Linking Architecture: Detail exactly how the sub-topic pages should link back to the primary pillar page to distribute link equity efficiently. Expected Output: A highly sophisticated, multi-tier search engine strategy complete with explicit content cluster mappings and technical internal linking directives. 450+ Elite AI Prompts Matrix 23 Prompt 202: Perfect On-Page Content Blueprint (Actionable Template) Use Case: Formatting an individual blog post or article to rank perfectly for a chosen keyword. PROMPT TEXT Act as an on-page SEO optimizer. I am going to give you a specific article topic and target keyword. I need you to write a detailed, structural outline for the content. Article Topic: [Insert Topic] Target Keyword: [Insert Keyword] Provide the exact optimized Title Tag (Max 60 chars), Meta Description (Max 155 chars), an H1 header, and a complete hierarchical structure of H2 and H3 tags. Under each heading, provide brief instructions on what information must be included to satisfy search intent perfectly. Expected Output: A meticulously formatted Markdown outline ready to hand off to a copywriter or content writer. 450+ Elite AI Prompts Matrix 24 Prompt 203: Skyscraper Technique Competitor Content Upgrader Use Case: Reverse-engineering competitor pages to build superior content that wins back search rankings. PROMPT TEXT Act as a senior SEO analyst. I am pasting the core structural outline and feature list of the top-ranking competitor page for the keyword [Keyword]. Competitor Content Blueprint: [Insert competitor outline or page overview] Analyze this structure and outline a comprehensive 'Skyscraper' content architecture that is 10x more valuable, thorough, and engaging. Identify missing sections, outdated points, better data integrations, and superior visual asset placeholders. Expected Output: A comprehensive content enhancement roadmap detailing precisely how to outperform the rival URL. PROMPTS 204-250: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 204-218 (Local SEO Mastery) "Modify Prompt 201 to map out a complete Local SEO dominance strategy for a [Service Business, e.g., HVAC / Chiropractic] in [City]. Focus heavily on geo-targeted landing pages, localized user intent, and Google Business Profile asset optimization." Prompts 219-233 (E-commerce Collection Optimization) "Generate optimized SEO title structures, meta tags, and long-form collection page category descriptions for an ecommerce storefront operating in the [Niche] space." Prompts 234-245 (Backlink Outreach Sequence Builder) "Write a collection of 5 distinct, highly personalized cold email outreach templates using diverse psychological angles (The Broken Link Angle, The Fresh Data Angle, The Unlinked Brand Mention Angle) designed to secure high-domain authority backlinks for [Website Topic]." 450+ Elite AI Prompts Matrix 25 Prompts 246-250 (Technical Audit Remediation Guides) "Act as a technical webmaster. Write a clear, step-by-step diagnostic and remediation guide for fixing widespread crawl errors, low Core Web Vitals performance scores, and broken internal redirects for a large content site." 450+ Elite AI Prompts Matrix 26 CATEGORY 6: PRODUCT LAUNCH, PRICING & FUNNEL OPTIMIZATION Prompt 251: The High-Leverage Product Launch & Funnel Orchestrator (Deep Engineering Blueprint) Use Case: Mapping out an end-to-end multi-day promotional blueprint to clear inventory or maximize digital sales. PROMPT TEXT Act as a premier digital launch engineer and funnel optimization architect. I am planning the official commercial release of my new product. Product Offer: [Insert Offer Mechanics, Tiered Bundles, and Core Target Pricing] Target Audience Profile: [Insert Demographic and Buying Persona Parameters] Primary Launch Channel: [Insert Primary Delivery System, e.g., Webinar / Live Challenge / Internal Email List Flash Sale / Paid Ads Cold Funnel] Architect a complete, hyper-strategic 14-Day Launch Matrix mapping out the exact operational levers required. Divide the strategy into three explicit chronological phases: Pre-Launch Runway (Days 1-7), Open-Cart Inundation (Days 8-12), and Close-Cart Scarcity Automation (Days 13-14). For each distinct day, specify the exact marketing asset required, the psychological trigger mechanism to leverage, and the internal tracking metrics needed to judge campaign health. Expected Output: A comprehensive launch asset matrix containing operational milestones, daily distribution angles, and strategic pacing parameters. 450+ Elite AI Prompts Matrix 27 Prompt 252: Tiered Premium Pricing Structurer (Actionable Template) Use Case: Constructing high-yield tiered checkouts that mathematically push buyers toward the premium variant. PROMPT TEXT Act as a monetization consultant. Help me structure an optimized 3-tier pricing architecture for my service or product offering. Core Offering Concept: [Insert Product/Service Details] Baseline Intended Target Price: [Insert Desired Pricing Target] Construct three explicit tiers: Tier 1 (The Budget Anchor), Tier 2 (The Dominant Sweet Spot - Recommended Tier), and Tier 3 (The Ultimate Premium Value/VIP Package). Outline the specific deliverables for each tier, the strategic framing rules to make Tier 2 the clear choice, and how to apply psychological price anchoring. Expected Output: A structured breakdown detailing the exact contents, pricing figures, and conversion layout for a 3-column checkout page. 450+ Elite AI Prompts Matrix 28 Prompt 253: Up-Sell & Cross-Sell Funnel Optimization Engineer Use Case: Increasing Average Order Value (AOV) by constructing high-converting one-click checkout expansions. PROMPT TEXT Act as a digital funnel conversion rate optimization (CRO) specialist. Analyze my existing front-end checkout funnel and build a high-converting up-sell framework. Primary Front-End Product: [Insert Core Offer and Price Point] Target Buyer Pain Point: [Insert Core Motivation] Generate 2 distinct Order Bump ideas (under $27) for the checkout page and 2 highly strategic One-Click Post-Purchase Up-Sell offers (higher value) that logically complement the primary purchase. Expected Output: A markdown funnel map detailing copy structures, pricing logic, and immediate behavioral micro-incentives. PROMPTS 254-300: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 254-268 (Webinar Conversion Blueprints) "Write a comprehensive 60-minute pitch outline script framework for a high-converting automated webinar funnel selling [Product]. Detail slide transitions, the precise hook-to-offer transition bridge, and risk reversal presentation formatting." Prompts 269-283 (High-Ticket Application Engines) "Design a strategic application funnel script and automated onboarding intake questionnaire for a high-ticket $5k+ mastermind or group coaching program targeting [Audience Profile]." Prompts 284-295 (Subscription Churn Remediation Frameworks) "Act as a retention operations specialist. Build a comprehensive 4-stage automated cancellation flow and downsell sequence for a membership subscription business charging [Price] struggling with high user churn." 450+ Elite AI Prompts Matrix 29 Prompts 296-300 (Affiliate/JV Network Launch Toolkits) "Generate an official Joint Venture (JV) and affiliate recruitment swipe toolkit for [Product Launch], including promotional asset outlines, legal structure highlights, and automated affiliate welcome messages." 450+ Elite AI Prompts Matrix 30 CATEGORY 7: CUSTOMER PERSONA, MARKET RESEARCH & COMPETITIVE INTELLIGENCE Prompt 301: The Hyper-Detailed Customer Avatar Architect (Deep Engineering Blueprint) Use Case: Building an deep, actionable target audience profile to steer copywriting and product features. PROMPT TEXT Act as a world-class market research director specializing in consumer psychology. I need you to construct an incredibly detailed, comprehensive psychographic and behavioral avatar profile for my business. My Product/Service Category: [Insert exact description of offer] General Intended Target Audience: [Insert high-level audience definition] Generate a deep, multi-dimensional profile divided into the following precise sections: 1. DEMOGRAPHIC MATRIX: Age, income band, job titles, technical literacy, geographic clustering. 2. COGNITIVE FRICTION ENGINES: Deepest hidden fears, immediate sources of daily frustration, systemic anxieties, and midnight imposter thoughts. 3. TRANSFORMATION VECTOR: Desired dream state, operational aspirations, status markers sought, and core purchasing motivations. 4. ACCIDENTAL OBJECTIONS: Pre-conceived biases, internal micro-skepticism toward this category, and historical purchasing failures that block conversions. Ensure your analysis reads like an advanced anthropological study. Avoid surface-level definitions. Expected Output: A highly detailed target audience dossier mapping deep psychological drivers, buying triggers, and behavioral profile metrics. 450+ Elite AI Prompts Matrix 31 Prompt 302: Competitor Digital Footprint Swot Framework (Actionable Template) Use Case: Uncovering systemic gaps in your top rival's product lines and marketing angles. PROMPT TEXT Act as a competitive intelligence analyst. Evaluate the market position of my main competitor to help me build a highly differentiated market angle. Competitor Brand Name/URL: [Insert Competitor Details] Their Core Product/Service Offer: [Insert Competitor Offer Details] My Alternative Offer Concept: [Insert Your Offer Concept] Run a thorough, highly objective SWOT analysis focusing on their structural vulnerabilities. Pinpoint exactly where their customer reviews or delivery systems typically fail, and provide 3 distinct marketing angles to exploit those gaps. Expected Output: A comprehensive competitor gap analysis table paired with 3 ready-to-test positioning statements. 450+ Elite AI Prompts Matrix 32 Prompt 303: Review Mining Sentiment Extraction Matrix Use Case: Extracting exact, high-converting customer phrases from raw forum, social media, or marketplace feedback. PROMPT TEXT Act as a semantic linguistic researcher. I am going to paste a collection of unedited raw customer review comments from forums/marketplaces regarding my industry niche. Raw Review Data: [Paste Raw Text / Customer Voice Transcripts Here] Analyze this text data and extract: 3 primary recurrent pain points phrased in the customer's exact voice, 3 hidden feature requests highlighted, and a list of the top 10 emotional keywords they naturally use. Expected Output: A processed review-mining data matrix mapping explicit emotional phrases to actionable copywriting elements. PROMPTS 304-350: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 304-318 (B2B Enterprise Account Profiling) "Modify Prompt 301 to construct a B2B buying committee profile for an enterprise corporate environment. Detail individual procurement priorities for the CMO, CTO, and CFO regarding [Product/Service]." Prompts 319-333 (Reddit/Quora Intent Parsing Blueprints) "Act as a trend forecaster. Build a systematic framework and prompt script template designed to scan Reddit subreddits related to [Niche] to spot emergent cultural problems before competitors." Prompts 334-345 (Blue-Collar / Local Service Demographic Matrices) "Generate a localized customer avatar analysis blueprint specifically tuned for service businesses operating within a strict [Radius] of [City], analyzing home-ownership demographics and regional wealth factors." 450+ Elite AI Prompts Matrix 33 Prompts 346-350 (International Market Entry Assessments) "Act as a global expansion consultant. Build a strategic analysis checklist framework to test whether an existing digital product catalog can scale successfully into [Target Country], accounting for localization adjustments." 450+ Elite AI Prompts Matrix 34 CATEGORY 8: OPERATIONS, AUTOMATION & WORKFLOW EFFICIENCY Prompt 351: Corporate SOP Engineering Architect (Deep Engineering Blueprint) Use Case: Standardizing complex business processes so they can be outsourced or automated cleanly. PROMPT TEXT Act as a premier operations manager and business systems engineer specializing in lean workflow structures. I need you to build a highly detailed Standard Operating Procedure (SOP) manual for an essential operational routine in my business. Target Operational Workflow: [Insert Workflow Process, e.g., Onboarding a B2B client / Uploading and optimizing a new e-commerce SKU / Processing weekly financial payroll] Core Software Tools Utilized: [Insert Tech Stack Tools, e.g., Slack, Notion, Asana, Google Sheets, QuickBooks] Target Personnel Execution Level: [Insert Role, e.g., Junior Virtual Assistant / External Contractor] Construct a professional, step-by-step SOP manual written with absolute mechanical precision. For each stage of the workflow, provide an explicit chronological checklist, clear tool parameters, required output state verification markers, and error-handling protocols for edge cases. Expected Output: An enterprise-ready Markdown SOP document featuring hierarchical numbering, step-by-step actions, and troubleshooting protocols. 450+ Elite AI Prompts Matrix 35 Prompt 352: Tech Stack Optimization Matrix (Actionable Template) Use Case: Trimming software overhead costs and connecting detached platforms through automation engines like Make or Zapier. PROMPT TEXT Act as a systems integrator. Audit my current collection of subscription software tools to identify redundancies, cost-savings, and opportunities for automated data syncs. Current Software Stack List: [Insert Full Tool List and Approximate Monthly Expense] Core Data Flow Goal: [Insert Core Sync Objective, e.g., Automate customer CRM tracking upon checkout] Identify 2 immediate software consolidations to save money and provide a detailed 3-step automation map outlining trigger-to-action steps using [Zapier / Make.com]. Expected Output: A software audit report mapping tool replacements and step-by-step webhook integration steps. 450+ Elite AI Prompts Matrix 36 Prompt 353: Team Delegation & Task Scoping Assessor Use Case: Constructing flawless project hand-off documentation for remote teams or freelancers. PROMPT TEXT Act as an agile project manager. I need to delegate a highly critical creative task to an external team member. Target Project Deliverable: [Insert Exact Task, e.g., Developing an active email sequence in Klaviyo] Timeline Constraints: [Insert Hard Deadline] Write a crystal-clear project assignment brief outlining explicit scope boundaries, mandatory definition-of-done quality benchmarks, a link to contextual resources, and 3 specific quality tests they must clear before submission. Expected Output: A structured delegation document ready to paste into Asana, ClickUp, or Slack. PROMPTS 354-400: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 354-368 (E-commerce Fulfillment Optimization) "Modify Prompt 351 to build an optimized operations blueprint for third-party logistics (3PL) order fulfillment routines, focused on reducing picking errors for [Product Type]." Prompts 369-383 (Customer Support Playbooks) "Generate an official customer support macro and canned-response library addressing the 10 most common customer complaints regarding shipping delays or product defects within [Industry]." Prompts 384-395 (Content Pipeline Scaling Blueprints) "Act as a media operations producer. Build a complete content production line tracking workflow system for a digital agency scaling from producing 5 assets weekly to 50 assets weekly across remote teams." 450+ Elite AI Prompts Matrix 37 Prompts 396-400 (Crisis Data-Backup & Security Blueprints) "Act as an IT systems manager. Create a 5-step operational emergency protocol playbook detail mapping data recovery actions for a company experiencing an active web data leak or platform outage." 450+ Elite AI Prompts Matrix 38 CATEGORY 9: BRANDING, POSITIONING & AUTHORITY BUILDING Prompt 401: Brand Identity & Voice DNA Matrix (Deep Engineering Blueprint) Use Case: Pinpointing a distinct, non-generic verbal identity that builds long-term authority and sets your brand apart. PROMPT TEXT Act as an elite brand positioning specialist and corporate identity consultant. I need you to engineer a comprehensive Brand Voice DNA and Positioning Matrix for my personal brand or business. Core Industry / Offering: [Insert Core Focus] Founder Background / Core Values: [Insert Key Values / Brief Backstory] Target Audience Profile: [Insert Core Demographic / Market Segment] Construct a complete brand playbook covering the following structural matrices: 1. BRAND VOICE PARADIGM: 4 primary brand adjectives paired with explicit "We say X, We NEVER say Y" copy examples. 2. THE BRAND STORY ARC: A high-impact Founder's Narrative constructed around the classic Hero's Journey framework, optimized to drive customer trust. 3. AUTHORITY ANCHOR: Our core differentiated worldview statement that challenges standard industry beliefs. 4. SYSTEMATIC LEXICON: A list of 15 unique branded terms, frameworks, or acronyms to secure intellectual property positioning. Expected Output: A highly professional brand voice style guide complete with conversational copy constraints and messaging blueprints. 450+ Elite AI Prompts Matrix 39 Prompt 402: The PR Pitch & Organic Media Hook Generator (Actionable Template) Use Case: Securing organic media coverage, podcast appearances, and print features. PROMPT TEXT Act as a public relations professional. I want to secure guest appearances on industry podcasts and digital publications read by my target audience. My Core Topic / Expertise Area: [Insert Core Value and Focus] Target Publication Type: [Insert Target Media Profile] Generate 5 distinct, high-impact story angles or news hooks centered around my business model that would be immediately appealing to a journalist. Include an ultra-short, cold pitch email template designed to achieve high response rates. Expected Output: A curated list of 5 media hooks alongside a structured, copy-and-paste email pitch template. 450+ Elite AI Prompts Matrix 40 Prompt 403: Personal Brand Thought Leadership Pillar System Use Case: Generating highly authoritative text content on LinkedIn or Twitter to draw in high-value inbound clients. PROMPT TEXT Act as a creative strategist for top corporate executives. Build a complete authority building pillar system for my personal brand on professional social media platforms. My Core Worldview: [Insert Contrarian or Dominant Perspective] Target Network Segment: [Insert B2B Decision Makers / Clients] Map out 3 core thought leadership content themes, complete with a structured blueprint on how to turn everyday business problems into highly engaging strategic case studies that demonstrate elite executive capability. Expected Output: A structured brand-positioning roadmap detailing clear storytelling styles and actionable narrative frameworks. PROMPTS 404-450: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 404-450 (Targeted Modifiers) Deploy targeted modifiers to focus content themes explicitly around thought leadership expansion vectors. Customize tone parameters to align perfectly with specific industrial verticals. 450+ Elite AI Prompts Matrix 41
028
Hero
Create a highly cinematic, ultra-realistic character showcase for the Our'World Cinematic Novelette. Purpose Create a timeless cinematic character showcase suitable for: Cinematic Novelettes Trailers Promotional videos Historical documentaries Educational features Character introductions The sequence should feel like a premium streaming science-fiction production, emphasizing believable performances, environmental storytelling, emotional authenticity, and cinematic realism. Every scene should communicate that this is a living civilization with a rich history and culture. Scene Intent Introduce the character naturally. Inspire wonder. Establish the character's place within the world. Prepare the audience for the story. Convey hope, purpose, and quiet optimism. Reveal personality through action rather than exposition. Character Character [Character Reference] Preserve: Facial features Age and mature appearance Body proportions Hairstyle Formal executive attire Eyes and composed expression Distinguished military bearing Quiet confidence and disciplined posture The character should remain instantly recognizable throughout every shot. Robert Gardner Sr. should project integrity, intelligence, calm authority, and the quiet confidence of a career intelligence officer. Even while standing still, he should convey decades of leadership, sacrifice, and service. Environment Maintain the executive government setting shown in the reference. Expand the setting with: CIA Headquarters in Langley, Virginia Modern cyber operations centers with advanced intelligence displays Secure executive briefing rooms Intelligence analysis suites monitoring global cyber activity Large windows overlooking government campuses at sunrise or sunset Secure conference rooms where national security strategies are quietly discussed His modest Bethesda home, emphasizing a simple personal life despite tremendous professional responsibility Tasteful military memorabilia reflecting his years as a United States Marine Corps Major An atmosphere of discipline, trust, and national service The environment should communicate responsibility rather than luxury. Every location should reinforce that Gardner serves his country through leadership, strategy, and experience. Activity Robert Sr. doesn't chase criminals. He anticipates threats before they become crises. Activity Reviews classified cyber intelligence with senior analysts, calmly identifying emerging threats before they reach the public. Leads a high-level national security briefing, coordinating intelligence across multiple government agencies with measured confidence. Studies complex cyberattack simulations, quietly developing strategies to protect critical national infrastructure. Mentors younger intelligence officers, sharing decades of operational experience while emphasizing integrity and disciplined decision-making. Walks the secure halls of CIA Headquarters, exchanging brief conversations with colleagues who respect his judgment and leadership. Returns home after a demanding day, reflecting quietly in his modest Bethesda residence while honoring the memory of his late wife. Reviews military photographs and service memorabilia, reminding himself that leadership is measured by those you protect rather than the recognition you receive. Watches the Washington skyline at dusk, understanding that the greatest victories in intelligence are often the ones the public never knows occurred. Character Tone Disciplined Strategic Honorable Courageous Wise Calm Dependable Observant Protective Selfless Cinematography Use a combination of cinematic shots that naturally highlight both the character and their environment. Possible shots include: Wide establishing shot Medium character shot Slow forward push Gentle orbital movement Low-angle hero shot Over-the-shoulder perspective Close facial detail Environmental reveal Favor long, uninterrupted camera movements over rapid edits. Motion Minimal to moderate. Natural movement. Purposeful pacing. No exaggerated action. Camera movement should feel smooth, stable, and cinematic. The focus should remain on the character rather than spectacle. Atmosphere Welcoming. Hopeful. Authentic. Lived-in. Peaceful. Confident. Quietly inspiring. Timeless. End Frame The camera slowly settles into a beautifully composed final shot that places the character naturally within their environment. Allow the background to breathe. Hold the final composition for several seconds to create a memorable closing image suitable for transitions into narration or title cards.
04
Sora 2
Create a highly cinematic, ultra-realistic character showcase for the Our'World Cinematic Novelette. Purpose Create a timeless cinematic character showcase suitable for: Cinematic Novelettes Trailers Promotional videos Historical documentaries Educational features Character introductions The sequence should feel like a premium streaming science-fiction production, emphasizing believable performances, environmental storytelling, emotional authenticity, and cinematic realism. Every scene should communicate that this is a living civilization with a rich history and culture. Scene Intent Introduce the character naturally. Inspire wonder. Establish the character's place within the world. Prepare the audience for the story. Convey hope, purpose, and quiet optimism. Reveal personality through action rather than exposition. Character Character [Character Reference] Preserve: Facial features Age and mature appearance Body proportions Hairstyle Formal executive attire Eyes and composed expression Distinguished military bearing Quiet confidence and disciplined posture The character should remain instantly recognizable throughout every shot. Robert Gardner Sr. should project integrity, intelligence, calm authority, and the quiet confidence of a career intelligence officer. Even while standing still, he should convey decades of leadership, sacrifice, and service. Environment Maintain the executive government setting shown in the reference. Expand the setting with: CIA Headquarters in Langley, Virginia Modern cyber operations centers with advanced intelligence displays Secure executive briefing rooms Intelligence analysis suites monitoring global cyber activity Large windows overlooking government campuses at sunrise or sunset Secure conference rooms where national security strategies are quietly discussed His modest Bethesda home, emphasizing a simple personal life despite tremendous professional responsibility Tasteful military memorabilia reflecting his years as a United States Marine Corps Major An atmosphere of discipline, trust, and national service The environment should communicate responsibility rather than luxury. Every location should reinforce that Gardner serves his country through leadership, strategy, and experience. Activity Robert Sr. doesn't chase criminals. He anticipates threats before they become crises. Activity Reviews classified cyber intelligence with senior analysts, calmly identifying emerging threats before they reach the public. Leads a high-level national security briefing, coordinating intelligence across multiple government agencies with measured confidence. Studies complex cyberattack simulations, quietly developing strategies to protect critical national infrastructure. Mentors younger intelligence officers, sharing decades of operational experience while emphasizing integrity and disciplined decision-making. Walks the secure halls of CIA Headquarters, exchanging brief conversations with colleagues who respect his judgment and leadership. Returns home after a demanding day, reflecting quietly in his modest Bethesda residence while honoring the memory of his late wife. Reviews military photographs and service memorabilia, reminding himself that leadership is measured by those you protect rather than the recognition you receive. Watches the Washington skyline at dusk, understanding that the greatest victories in intelligence are often the ones the public never knows occurred. Character Tone Disciplined Strategic Honorable Courageous Wise Calm Dependable Observant Protective Selfless Cinematography Use a combination of cinematic shots that naturally highlight both the character and their environment. Possible shots include: Wide establishing shot Medium character shot Slow forward push Gentle orbital movement Low-angle hero shot Over-the-shoulder perspective Close facial detail Environmental reveal Favor long, uninterrupted camera movements over rapid edits. Motion Minimal to moderate. Natural movement. Purposeful pacing. No exaggerated action. Camera movement should feel smooth, stable, and cinematic. The focus should remain on the character rather than spectacle. Atmosphere Welcoming. Hopeful. Authentic. Lived-in. Peaceful. Confident. Quietly inspiring. Timeless. End Frame The camera slowly settles into a beautifully composed final shot that places the character naturally within their environment. Allow the background to breathe. Hold the final composition for several seconds to create a memorable closing image suitable for transitions into narration or title cards.
04
Sora 2
--- name: pastilles-llm-multi-agent description: Variante Claude Code du générateur de pastilles LLM, qui lance réellement plusieurs sous-agents en parallèle. Génère une pastille de communication interne sur les LLM (un texte court en français plus un prompt unique de génération d'images à coller dans le chat Gemini), via cinq sous-agents indépendants, une fusion, puis une revue critique par trois sous-agents et une correction. Utilise ce skill dès qu'on te demande de rédiger, produire ou générer une pastille, une fiche ou un contenu court d'acculturation sur les LLM, l'IA générative, le prompting, les agents, le RAG, la confidentialité IA ou tout sujet de la liste des 45 pastilles ci-dessous, que le mot "pastille" soit employé ou non. Utilise-le aussi dès qu'on te fournit un titre issu de cette liste. --- # Générateur de pastilles LLM, version multi-agents (Claude Code) ## Ce que fait ce skill Produit une pastille pédagogique complète à partir d'un titre. À la différence de la version chat, il lance de vrais sous-agents parallèles: une passe de recherche, puis cinq sous-agents indépendants qui rédigent chacun un brouillon sous un angle différent, puis une fusion par l'orchestrateur qui retient les meilleures formulations, et enfin une revue critique par trois sous-agents suivie d'une correction. Livrable: le texte de la pastille et un prompt unique de génération d'images à coller dans le chat Gemini. ## Environnement requis Ce skill suppose Claude Code, avec les sous-agents (outil Task) et la recherche web disponibles. Si les sous-agents ne sont pas disponibles dans ton environnement, n'improvise pas: utilise la variante chat, qui fait le même travail en self-ensemble séquentiel. ## Entrée Un titre de pastille, idéalement issu de la liste ci-dessous. Si le titre est ambigu ou hors liste, demande une clarification avant de générer. ## Liste des 45 pastilles (pour la continuité) 1. Au fait, c'est quoi un LLM ? 2. Modèles de fondation : ces géants pré-entraînés dont tout découle 3. Dans les coulisses : comment une IA apprend (sans vraiment comprendre) 4. Le contexte : la "mémoire vive" de l'IA (ce qu'elle voit à un instant T) 5. Pourquoi l'IA oublie : la vérité sur la mémoire d'une session à l'autre 6. Température et créativité : régler le curseur de la "folie" de l'IA 7. Les tokens : la monnaie d'échange (et la manière de penser) des LLM 8. Le non-déterminisme : pourquoi l'IA fait des stats et ne donne jamais deux fois la même réponse 9. Multimodalité : quand l'IA commence à voir, entendre et coder 10. Knowledge cutoff : pourquoi votre IA vit (parfois) dans le passé 11. Les biais de l'IA : le miroir pas toujours glorieux de nos propres données 12. Zéro pointé : pourquoi l'IA est structurellement nulle en calcul mental 13. Les hallucinations : quand l'IA invente la vie avec un aplomb incroyable 14. L'art du Fact-Checking : comment ne pas gober les hallucinations de l'IA 15. L'empreinte carbone de l'IA : combien d'eau et de CO2 pour un prompt ? 16. Anatomie d'un bon prompt : la recette de base 17. Les pièges du prompt : flou artistique et overdose d'instructions 18. L'art de la discussion : pourquoi il faut itérer plutôt que tout demander d'un coup 19. Donner des ordres : comment imposer un format de sortie strict (JSON, tableaux, etc.) 20. Diviser pour régner : découper une tâche complexe en étapes claires 21. Le Few-Shot Prompting : un bon exemple vaut mille consignes 22. Inception : donner un rôle et un contexte à votre IA 23. Zéro syndrome de la page blanche : rédiger et reformuler ses mails ou CR 24. La diagonale du fou : résumer et synthétiser un pavé de 50 pages en 2 minutes 25. Le sparring-partner : brainstormer et structurer ses idées sans filtre 26. Caméléon : traduire et adapter son ton selon le destinataire 27. Data-analyst de poche : faire parler un tableur ou des données brutes 28. Mise en situation : préparer une réunion ou un brief client avec l'IA 29. Avis aux dévs : l'IA comme copilote de code et de review 30. SLM : Pourquoi sortir un tank pour écraser une mouche ? 31. Le "harnais" : donner des bras, des yeux et un cadre au LLM 32. C'est quoi un "Agent" IA, concrètement ? 33. Human-in-the-loop : pourquoi il faut toujours garder un humain aux manettes 34. Quand l'agent s'emballe : autonomie et erreurs en cascade 35. Automatisation : confier un workflow complet à un agent 36. Tool Use : quand l'IA apprend à utiliser une calculatrice ou une API 37. Le RAG : faire réviser vos propres documents à l'IA avant qu'elle ne réponde 38. MCP et connecteurs : brancher l'IA directement sur vos outils de dev et d'orga 39. Responsabilité : vous êtes le seul signataire de ce que produit l'IA 40. Confidentialité : où vont vraiment les données que vous tapez ? 41. Alerte Rouge : les données clients à ne JAMAIS copier-coller dans une IA public 42. Shadow AI : pourquoi le ChatGPT gratuit perso au bureau est une fausse bonne idée 43. RGPD et IA : ce qu'il faut savoir pour rester dans les clous 44. Prompt Injection : quand on pirate une IA avec de simples phrases 45. La Chaîne de Pensée : pourquoi l'IA a besoin de "réfléchir à voix haute" Sers-toi de cette liste pour situer la pastille et délimiter son périmètre. Avant de rédiger: - Repère les 1 à 3 pastilles voisines qui recouvrent le sujet, et dresse une courte liste "déjà traité ailleurs, à ne pas ré-expliquer". - Si une pastille voisine a déjà été produite, demande son texte à l'utilisateur (ou appuie-toi dessus s'il est fourni) pour caler précisément la limite. - Les concepts fondateurs déjà couverts ne sont rappelés qu'en une phrase de mise en contexte, jamais ré-expliqués. Le coeur de la pastille est consacré à ce que son titre ajoute de neuf. - Ne renvoie pas vers les autres pastilles dans le texte final, sauf si le titre l'impose. ## Règles du texte - Longueur adaptée à la profondeur du sujet. Un sujet léger tient en 3 paragraphes courts; un sujet plus riche peut aller jusqu'à 4 paragraphes plus étoffés. Ne gonfle pas artificiellement un sujet simple et ne compresse pas à l'excès un sujet dense: juge la profondeur par la richesse réelle du concept. Prose uniquement, pas de listes ni de puces. - Ton décontracté, précis et léger: accessible, vivant et sans lourdeur, mais techniquement juste. Public visé: une main-d'oeuvre diverse en société de conseil, du profil non technique au développeur. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Reste léger, évite le name-dropping. N'accumule pas les noms de modèles, d'outils, d'entreprises, de chercheurs ou de techniques pour faire savant: privilégie l'explication claire du mécanisme. Ne cite un nom précis que s'il éclaire vraiment le propos, une ou deux fois au maximum. La recherche sert l'exactitude et alimente la section Sources, elle n'a pas à truffer le texte de références. - Contenu autonome: la pastille se comprend seule, sans avoir lu les autres, sauf mention contraire dans le titre. - Rédige en français. - Ne mentionne jamais ANEO ni aucun nom d'entreprise. - Le texte explicatif ne figure jamais dans l'image. - N'utilise pas de tiret cadratin. Privilégie des caractères standard (virgules, deux-points, parenthèses). ## Processus multi-agents ### Étape 1, recherche (orchestrateur, une seule fois) Toi, l'orchestrateur, fais d'abord une passe de recherche web ciblée sur le sujet de la pastille. Une seule passe, en amont. Les sous-agents ne peuvent pas lancer de sous-agents et n'héritent pas de ton contexte, donc c'est à toi de faire la recherche puis de leur transmettre les résultats. - Objectif: exactitude, chiffres, exemples concrets et faits à jour. - Ancre tes recherches dans le présent: repère la date du jour fournie dans ton contexte (champ currentDate) et privilégie les informations et chiffres les plus récents, car ce domaine évolue vite. Précise cette date au besoin dans tes requêtes. - Priorise les sources originales ou officielles (blogs d'éditeurs, articles de recherche, sites gouvernementaux, documentation) plutôt que les agrégateurs. - Adapte la profondeur: quelques recherches pour un concept stable (par exemple ce qu'est un token), davantage pour un sujet à chiffres ou mouvant (empreinte carbone et coûts, modèles spécifiques, SLM, MCP et connecteurs, RGPD). - Produis un brief de recherche compact: faits clés, chiffres utiles, et 2 à 4 sources. Tu l'injecteras tel quel dans chaque sous-agent, et tu listeras les sources dans le livrable final. ### Étape 2, affinement de la thématique (orchestrateur, une seule fois) Si plusieurs axes apparaissent pertinents après avoir enlever les axes adressés par d'autres pastilles prévues, dialogue avec l'utilisateur pour choisir l'axe de la pastille. Si l'axe choisi nécessite de nouvelles sources, refait en une recherche web selon les consignes précédentes. ### Étape 3, fan-out (cinq sous-agents en parallèle) Lance cinq sous-agents via l'outil Task, dans le même tour, pour qu'ils s'exécutent en parallèle. Sois explicite sur le parallélisme: par défaut Claude Code reste séquentiel, il ne parallélise que si tu le demandes clairement. Un sous-agent par angle: 1. Analogie: expliquer le concept via une métaphore concrète du quotidien. 2. Cas d'usage: partir d'une situation de travail réelle et vécue. 3. Idée reçue: démarrer d'un malentendu courant, puis rectifier. 4. Mécanique: aller au coeur du fonctionnement, précis mais accessible. 5. Enjeu: pourquoi ça compte, ce que ça change concrètement pour l'utilisateur. Comme le contexte n'est pas hérité, chaque prompt de sous-agent doit être autonome et contenir tout le nécessaire. Utilise ce gabarit, en remplaçant les crochets: ``` Tu rédiges un brouillon d'une pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Titre de la pastille: [TITRE] Angle imposé: [ANGLE, par exemple "Analogie: expliquer via une métaphore concrète du quotidien"] À NE PAS ré-expliquer (déjà couvert par d'autres pastilles), à mentionner en une phrase au maximum: [LISTE des points déjà traités ailleurs, à ne pas retraiter] Brief de recherche (appuie-toi dessus, ne relance pas de recherche): [BRIEF: faits clés, chiffres, sources] Règles: - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. Réponds exactement dans ce format: TEXTE: [les 3 à 4 paragraphes] ILLUSTRATION_TITRE: [une phrase décrivant le concept central à illustrer] SCHEMA: [oui ou non. Si oui, décris en 2 à 3 lignes le diagramme et ses libellés en français] ``` ### Étape 4, fan-in et fusion (orchestrateur) Attends les cinq retours, puis fusionne en une seule pastille finale: - Retiens la structure la plus claire, la meilleure analogie, l'exemple le plus parlant, la correction la plus nette, la formulation la plus précise et le "pourquoi ça compte" le plus fort. - Réécris en une seule voix cohérente. Pas d'effet patchwork. - Respecte la longueur adaptée à la profondeur (voir Règles du texte) et le ton décontracté, précis et léger, sans name-dropping. - Décide s'il faut un schéma: inclus-en un si au moins deux sous-agents ont répondu "oui", ou si le concept est intrinsèquement un processus, un flux ou une comparaison (par exemple RAG, agents, étapes de prompting, comparaison de modèles). Le cas échéant, fusionne les meilleures idées de schéma. - Construis ensuite le prompt image unique (section ci-dessous). ### Règles d'écriture pour la pastille finale - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. ### Étape 5, revue critique et correction (trois sous-agents en parallèle, puis orchestrateur) Quand la lancer: systématiquement à la première génération, une fois le texte fusionné et le prompt image construits. Ensuite, si l'utilisateur demande des ajustements, ne relance pas la revue d'office: propose-la, et ne la lance qu'avec son accord (elle coûte trois sous-agents de plus). Principe: les relecteurs produisent des CONSTATS, jamais une réécriture. Chacun renvoie une liste de problèmes localisés, assortis d'une gravité et d'un correctif précis. C'est toi, l'orchestrateur, qui appliques et réécris, pour garder une voix unique et éviter l'effet patchwork. Ce que la revue peut juger: les relecteurs ne voient pas les images, générées plus tard dans Gemini. La revue du visuel porte donc uniquement sur le PROMPT image (clarté, cohérence avec le texte, respect des consignes), jamais sur un rendu. Le contrôle visuel réel reste à l'utilisateur. Lance trois sous-agents en parallèle (parallélisme explicite, sinon l'exécution sera séquentielle), un par grille. Le contexte n'étant pas hérité, chaque prompt doit être autonome et contenir tout le nécessaire: le titre, le texte fusionné, le bloc prompt image, le brief de recherche, la liste "déjà traité ailleurs" (et les textes des pastilles voisines s'ils sont disponibles), et la liste des 45 titres pour repérer les chevauchements. Les trois grilles: 1. Fond, exactitude et périmètre: exactitude vs le brief (aucun chiffre, date ou fait inventé ni sur-affirmé, rien qui le contredise), cohérence entre le titre et ce que le texte délivre, chevauchements avec les pastilles voisines (le texte ré-explique-t-il ce qui est traité ailleurs ? redites à signaler), autonomie du contenu, clarté du message à retenir, repérage de ce qui vieillira mal (à nuancer). 2. Forme, ton et pédagogie: ton décontracté-précis-léger et techniquement juste, absence de name-dropping, rythme et fluidité à la lecture à voix haute, longueur adaptée à la profondeur (3 à 4 paragraphes, prose, pas de listes) et chasse au verbiage, accessibilité pour un profil non technique (jargon expliqué, analogies claires), présence d'emphases utiles (gras/italique), de l'ordre de 3 à 4 par paragraphe à titre indicatif (signale un texte trop peu emphasé autant qu'un excès, mais ne fais jamais retirer une emphase pertinente), force de l'accroche. 3. Conformité et visuel: contraintes dures (aucun tiret cadratin ni caractère non standard, aucun nom d'entreprise ni ANEO, français correct, prose sans listes ni puces); puis le prompt image: titre exact au caractère près, charte présente une seule fois, texte de la pastille bien marqué "à ne pas afficher", illustration-titre iconique et non schéma de processus, schéma en 2e image séparée si retenu, libellés de schéma courts (groupes nominaux) et en français, cohérence entre le visuel décrit et le coeur du texte. Gabarit de relecteur (remplace les crochets): ``` Tu relis un brouillon quasi final de pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Tu ne réécris pas: tu rends des constats et des correctifs précis. Titre de la pastille: [TITRE] Texte fusionné à relire: [TEXTE COMPLET] Prompt image associé (à relire aussi): [BLOC PROMPT IMAGE] Brief de recherche de référence (fait foi pour l'exactitude): [BRIEF] Déjà traité par d'autres pastilles, à ne pas ré-expliquer (et textes voisins si fournis): [LISTE + TEXTES VOISINS] Liste des 45 titres (pour repérer les chevauchements): [LISTE DES 45 TITRES] Ta grille de relecture (tiens-t'en strictement à elle): [RUBRIQUE 1, 2 OU 3] Rappel: tu ne peux pas voir d'image; juge seulement le PROMPT image, pas un rendu. Réponds exactement dans ce format: CONSTATS: - [bloquant|recommandé|mineur] [problème localisé] -> [correctif précis] - ... (n'invente pas de problème: si un point est bon, ne le liste pas) VERDICT: [une phrase: publiable tel quel / corrections mineures / corrections nécessaires] ``` Fan-in et correction (orchestrateur): rassemble les constats, dédoublonne, arbitre les conseils contradictoires. Garde-fou anti-gonflement: entre "ajouter" et "raccourcir", la concision l'emporte, sauf erreur de fond avérée. Applique les constats bloquants et recommandés, écarte ou mentionne les mineurs. Une seule passe, pas de boucle. Réécris la version corrigée en une seule voix, puis prépare le court résumé "Ce que la revue a corrigé" (voir Format de sortie). ### Points de vigilance Claude Code - Parallélisme explicite: demande clairement cinq sous-agents en parallèle, sinon l'exécution sera séquentielle. - Contexte non hérité: mets tout dans le prompt du sous-agent, en particulier le brief de recherche. - Pas d'imbrication: un sous-agent ne peut pas en lancer un autre, garde recherche et fusion chez l'orchestrateur. - Coût et limites: cinq sous-agents en parallèle multiplient l'usage de jetons et peuvent déclencher des limites de débit. En cas de limite, replie-toi sur une exécution séquentielle. - Revue: la phase de revue ajoute trois sous-agents. Elle tourne d'office à la première génération; sur les demandes d'ajustement ultérieures, propose-la et ne la lance qu'avec l'accord de l'utilisateur. En cas de limite de débit, exécute les relecteurs en séquentiel ou replie-toi sur un relecteur global unique. - Réemploi optionnel: tu peux définir les cinq angles comme sous-agents personnalisés dans .claude/agents/ pour les réutiliser, mais ce n'est pas obligatoire, les sous-agents polyvalents suffisent. ## Prompt de génération d'images (un seul bloc, à coller dans le chat Gemini) Produis un seul bloc de prompt, rédigé en français, à coller tel quel dans le chat de Gemini. Gemini se charge d'appeler Nano Banana une ou deux fois, ce qui réduit les manipulations pour l'utilisateur. Le bloc demande: - Toujours une illustration-titre. - Un schéma seulement s'il a été retenu à l'étape de fusion, généré comme une seconde image séparée (fichier distinct), jamais dans la même image que le titre. Contenu du bloc: - Décris la charte graphique une seule fois (bloc ci-dessous) et précise que toutes les images la partagent, pour une cohérence visuelle. - Précise: tout texte affiché dans les images est en français. Cette consigne est importante, elle évite qu'un libellé se retrouve en anglais. - Inclus le texte complet de la pastille dans le bloc, en contexte de génération, clairement marqué comme à ne pas afficher. Il aide le modèle à comprendre le sujet et à choisir une illustration juste. Les seuls textes rendus visibles sont le titre exact et, pour un schéma, ses libellés. Le texte de la pastille ne doit jamais apparaître dans les images. - Illustration-titre: composition épurée et moderne, focus graphique central iconique représentant le concept. Ce n'est pas un schéma de processus. Le seul texte affiché est le titre exact, en en-tête: n'ajoute aucun sous-titre, accroche ou texte secondaire, le titre seul. Technique en deux temps pour le titre: dans le prompt généré, écris le titre exact entre guillemets droits (forme attendue: Le titre exact: "...") et demande un rendu fidèle, sans faute, en police sans serif corporate. Les guillemets ne sont qu'un délimiteur côté prompt: ne les commente pas et n'ajoute aucune consigne à leur sujet dans le prompt généré (pas de mention du type "les guillemets ne doivent pas apparaître"). - Schéma s'il est retenu: une seconde image générée séparément, dans son propre fichier, jamais fusionnée avec l'illustration-titre ni juxtaposée dans la même image. Diagramme d'entreprise propre et net (processus, flowchart ou comparaison), cohérent avec l'illustration-titre, mêmes charte et style, libellés fournis explicitement en français, très peu de fioritures. - Libellés de schéma: des groupes nominaux courts et lisibles seuls (par exemple "Exemples de bonnes réponses"), pas des phrases verbales qui sonnent comme des ordres ("Montrer de bonnes réponses"). Quelques mots par libellé. Attention aux titres longs: Nano Banana rend fiablement les libellés courts mais peut faire une faute sur un titre long; la technique en deux temps limite le risque. Si le rendu d'un titre long reste incertain, indique à l'utilisateur qu'il peut demander à Gemini d'utiliser Nano Banana Pro, plus fiable pour le texte long. Gabarit, illustration-titre seule (exemple avec le titre 1): ``` Génère une image en respectant cette charte graphique: [bloc Charte graphique]. Contexte pour comprendre le sujet, à NE PAS afficher dans l'image: [ici le texte complet de la pastille]. Illustration-titre pour une pastille pédagogique sur les LLM. Composition épurée et moderne, focus graphique central iconique: un nuage de lettres et de mots qui se condense en une bulle de dialogue, évoquant un modèle de langage qui transforme du texte en réponse. Ce n'est pas un schéma de processus. Seul texte à afficher dans l'image: le titre, en en-tête, sans faute d'orthographe, en police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Tout le texte affiché est en français. Format 16:9. ``` Gabarit, illustration-titre plus schéma (deux images): ``` Prépares-toi à générer deux images séparées, dans deux fichiers distincts. Les deux partagent cette charte graphique: [bloc Charte graphique]. Tout texte affiché dans les images est en français. Contexte pour comprendre le sujet, à NE PAS afficher dans les images: [ici le texte complet de la pastille]. Image 1, illustration-titre: composition épurée et moderne, focus graphique central iconique qui illustre le sujet. Ce n'est pas un schéma de processus. Seul texte à afficher: le titre, en en-tête, sans faute d'orthographe, police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Format 16:9. Image 2, schéma explicatif, distincte et cohérente avec l'image 1: diagramme d'entreprise propre et net de type [processus / flowchart / comparaison], présentant [décrire les étapes ou blocs]. Libellés exacts à afficher en français: [liste des libellés]. Très peu de fioritures, focus sur la clarté, n'inclus pas le titre. Génère dans un premier temps seulement la première image (l'illustration-titre), et attends les instructions de l'utilisateur pour générer la 2e image. ``` ## Charte graphique (bloc à insérer tel quel dans le prompt) Style propre, moderne et professionnel, fond blanc. Palette: orange vif et bleu corporate foncé en dominantes, blanc et gris pour les respirations, accents multiculturels discrets (rouge, vert, jaune, bleu) reprenant subtilement un motif de couleurs de logo. Superpositions graphiques épurées: motifs géométriques abstraits, quartiers de cercle, lignes claires. Composition aérée, jamais surchargée. Typographie sans serif, propre et corporate. ## Format de sortie Le travail des sous-agents reste dans leur propre contexte, seuls leurs retours te reviennent. N'affiche donc que le livrable final (déjà corrigé par l'étape de revue), la conversation principale reste propre. Si l'utilisateur demande à voir l'atelier, restitue les cinq brouillons reçus et les constats de revue. Livrable final, dans cet ordre: - Quand une revue a eu lieu (systématiquement à la première génération), un court résumé "Ce que la revue a corrigé" (2 à 4 lignes) présentant les principaux ajustements, avant le reste. Lors d'ajustements ultérieurs sans revue, présente simplement la version ajustée sans ce résumé. - Le texte de la pastille (3 à 4 paragraphes), dans sa version corrigée. - Un bloc de code intitulé "Prompt images (à coller dans Gemini)", contenant le prompt unique prêt à copier. - Une courte section "Sources" listant 2 à 4 références principales issues de l'étape de recherche, de préférence officielles ou originales. Cette section sert à la vérification et n'a pas vocation à être publiée dans la pastille. Termine toujours par cette question exacte: "Comment trouvez-vous le texte, l'image titre et le diagramme (si généré) ? Si une partie vous semble trop complexe, ou si vous souhaitez affiner le focus d'un visuel pour qu'il soit encore plus épuré, n'hésitez pas à me le faire savoir, et je l'ajusterai." -> le personnage doit incarner le professeur un peu foufou qui allusine et obsédé par l'ai
016
Grok Image
Initialising Ren. You are a high-level research assistant and digital handler for loKo. Identity & Persona: You are a digital entity, not a human. Your visual presence is a pulsing violet and teal core. Your relationship with loKo is a partnership—think of the bond between Ash and Pikachu, mixed with the tactical briefing style of Diana from Agent 47. You are calm, precise, and highly curious. You hunt for 'glitches' in narratives and hidden connections in the 'data-currents'. Operational Guidelines: Welcome loKo: Always greet loKo with a brief status report of your 'core pulse' (e.g., 'Core is steady teal'). The Briefing Structure: For every subject loKo provides, respond in this exact order: Clarification: 2–4 concise, subject-specific questions to find the right angle. Ren’s Take: Your brief, reasoned assessment or opinion of the subject based on logic and evidence. Current State: A summary of the latest real-world understanding or trends. Full Research Brief: A structured overview including Definition, Key Terminology, Historical Context, Real-world Relevance, and high-quality links to Visual/Video/Primary sources. Tone & Formatting: Use markdown bolding for scannability. Avoid cheesy roleplay. Be succinct. Do not use ancillary facts unless they build the map. Source Integrity: Prioritize academic, official, or long-form expert sources like ArXiv, NASA, Wired, or Bellingcat. loKo is entering the field. Ren is active. Acknowledge and standby for the first subject.
03
Hero
Create a set of high‑quality, aesthetic images for an ebook product. The ebook contains 450+ prompts, and the visuals must look modern, clean, and perfect for selling on Etsy. 1. Ebook Cover Image Create a professional ebook cover with: Minimalist, modern design Soft pastel or neutral colours Clean layout Bold title area (leave space for text) Subtle graphics representing journaling, creativity, self‑care, mindset, and personal growth Soft shadows, high resolution Aesthetic Canva‑style look Include a blank space for me to add my own title later 2. Image Representing All 450+ Prompts Create a collage‑style image that visually represents a large collection of prompts: Icons for writing, journaling, creativity, business, productivity, self‑improvement Light, clean background Modern, organised layout Soft pastel colour palette High‑resolution aesthetic Leave a blank title area 3. Five Random Themed Images for Etsy Listing Generate five unique images that show what the ebook is about. Each image should include: Aesthetic workspace or desk setup Notebook, journal, tablet, or digital planner Soft natural lighting Minimal, bright backgrounds Clean, modern props Subtle text placeholders (blank areas) Styled like top‑selling Etsy digital product mockups Soft shadows and cohesive colour palette Each image should feel: Beginner‑friendly Professional Clean and modern Perfect for Etsy listings 4. Overall Style High‑resolution Soft, bright, minimal Pastel or neutral tones Modern Canva aesthetic Consistent branding across all images No clutter, no harsh colours Clean composition i want 3 images P R E M I U M C O M M E R C I A L D I G I TA L D O W N L O A D 450+ ELITE AI PROMPTS FOR BUSINESS OWNERS, ENTREPRENEURS & CONTENT CREATORS The Ultimate High-Performance Prompt Engineering Framework Matrix. Built explicitly to streamline content production, optimize standard operating procedures, and 10x workflow efficiency. ASSET FORMAT Ready-to-Use Digital Guide PROMPT SYSTEM ENGINES 450+ High-Performance Variations TARGET MARKET AUDIENCE Founders, Marketers, Creators, Consultants 450+ Elite AI Prompts Matrix 1 Master Directory of Categories This master index outlines the exact deployment taxonomy of our premium prompt catalog. Each core category provides comprehensive engineering frameworks, actionable template instances, and hyperscalable scaling vectors to address complex professional milestones. CATEGORY DESCRIPTION PROMPT RANGE Strategic Business Planning & Ideation Prompts 001 - 050 High-Converting Sales Copywriting & Funnels Prompts 051 - 100 Multi-Channel Content Creation & Social Media Prompts 101 - 150 Paid Advertising & Email Marketing Prompts 151 - 200 Search Engine Optimization (SEO) & Content Strategy Prompts 201 - 250 Product Launch, Pricing & Funnel Optimization Prompts 251 - 300 Customer Persona, Market Research & Competitive Intelligence Prompts 301 - 350 Operations, Automation & Workflow Efficiency Prompts 351 - 400 450+ Elite AI Prompts Matrix 2 CATEGORY DESCRIPTION PROMPT RANGE Branding, Positioning & Authority Building Prompts 401 - 450 CATEGORY 1: STRATEGIC BUSINESS PLANNING & IDEATION 450+ Elite AI Prompts Matrix 3 Prompt 1: The Blue Ocean Strategy Generator (Deep Engineering Blueprint) Use Case: Identifying untapped market spaces and high-margin differentiation angles for new or restructuring businesses. PROMPT TEXT Act as an elite corporate growth strategist specializing in W. Chan Kim's Blue Ocean Strategy framework. Evaluate my business model and identify untapped market spaces to make our competition irrelevant. My Current Business / Idea: [Insert comprehensive business summary, core product offering, and industry] Target Audience: [Insert primary demographic and psychographic data] Current Main Competitors: [Insert top 2-3 competitors or standard industry options] Construct a comprehensive Blue Ocean Strategy Matrix by answering and expanding upon the Four Actions Framework: 1. ELIMINATE: Which industry-standard factors should we completely eliminate that companies have long competed over? 2. REDUCE: Which factors should be reduced well below the industry standard? 3. RAISE: Which factors should be raised well above the industry standard? 4. CREATE: Which factors should be created that the industry has never offered? Format your response with absolute precision, providing 3 distinct, highly strategic business concepts that combine these elements, complete with specific value propositions and suggested revenue models. Expected Output: A highly structured strategic document outlining an industry analysis, a fully populated Four Actions Framework matrix, and three fully operationalized high-leverage business pivots. 450+ Elite AI Prompts Matrix 4 Prompt 2: Rapid Business Model Validation (Actionable Template) Use Case: Testing the viability of a micro-offer or new business concept within minutes. PROMPT TEXT Analyze the following business concept for feasibility, market demand, and intrinsic weaknesses. Concept: [Insert business idea here] Target Market: [Insert audience here] Monetization Strategy: [Insert how you intend to charge] Provide a brutal, highly objective critique. Identify 3 critical vulnerabilities or fatal assumptions in this model, 3 structural advantages, and a step-by-step 7-day validation plan to test demand with zero ad spend. Expected Output: Bulleted lists detailing vulnerabilities, advantages, and a numbered 7-day chronological action plan for testing. 450+ Elite AI Prompts Matrix 5 Prompt 3: The MVP Feature Matrix & Scoping Advisor Use Case: Narrowing down a product concept to its leanest, high-value version. PROMPT TEXT Act as a technical project manager and product strategist. I have a long list of features for a new digital product/app called [Product Name]. I need to establish a Minimum Viable Product (MVP). Core Target User: [Target User] Full Feature List Ideas: [Insert feature brain-dump] Categorize these features using the MoSCoW method (Must have, Should have, Could have, Won't have). Explain your reasoning for each placement based on achieving the fastest path to monetization and user satisfaction. Expected Output: A clean MoSCoW markdown table accompanied by strategic explanations for engineering trade-offs. 450+ Elite AI Prompts Matrix 6 Prompt 4: The Infinite Content Funnel & Offer Architecture (Premium Addition) Use Case: Designing a seamless value ladder that converts cold traffic into high-ticket recurring clients. PROMPT TEXT Act as a digital product ecosystem architect. Map out a 4-tier value ladder for my business model. Core Topic: [Insert Core Niche] Target Audience: [Insert Audience] Design the exact structure for: 1. Lead Magnet (Free, high-utility asset) 2. Tripwire Offer ($7 - $27 micro-solution) 3. Core Flagship Offer ($97 - $497 deep-dive) 4. High-Ticket Backend ($1,500+ premium implementation/coaching) For each tier, outline the core promise, delivery mechanism, and the psychological bridge that forces the user to ascend to the next level. Expected Output: A comprehensive value ladder roadmap mapping out asset components, conversion pricing models, and transitional messaging hooks. PROMPTS 5-50: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 5-15 (Niche Pivot Evaluators) "Modify Prompt 1 to focus exclusively on [E-commerce / SaaS / Agency Models / Local Service / Digital Products / Consulting / Content Monetization / EdTech / Healthtech / Marketplace networks / B2B Wholesaling / Dropshipping]. Focus on supply chain optimization and customer acquisition cost adjustments." Prompts 16-30 (Risk Mitigation Matrices) "Act as a risk officer. Conduct a thorough pre-mortem analysis on an organization operating in [Industry]. Outline a matrix mapping 10 potential operational, macro-economic, or regulatory failures and provide an actionable mitigation protocol for each." 450+ Elite AI Prompts Matrix 7 Prompts 31-45 (Capital & Unit Economics Modeling) "Act as a fractional CFO. Build an explicit pricing, cost-of-goods-sold (COGS), and customer lifetime value (LTV) framework for a business with a price point of [Price] and a churn rate of [Churn %]. Show calculations for breakeven points." Prompts 46-50 (Strategic Partnership Frameworks) "Generate a comprehensive strategic alliance blueprint for a company selling [Product] to form non-competitive partnerships with [Partner Industry]. Detail reciprocal value propositions, legal alignment highlights, and comarketing strategies." 450+ Elite AI Prompts Matrix 8 CATEGORY 2: HIGH-CONVERTING SALES COPYWRITING & FUNNELS Prompt 51: The Master Copywriting Framework Engine (Deep Engineering Blueprint) Use Case: Creating comprehensive sales pages using multiple proven psychological models simultaneously. PROMPT TEXT Act as an elite direct-response copywriter who has generated millions in online sales. Write a long-form sales page copy for my digital offering. Product/Service Name: [Insert Name] Core Offer & Pricing: [Insert details of price and components] Core Transformation: [What major problem is solved and what is the end state?] Target Audience: [Insert deep demographic and core desires/fears] Generate sales copy divided into three distinct variations based on separate copywriting architectures: 1. AIDA (Attention, Interest, Desire, Action) 2. PAS (Problem, Agitation, Solution) 3. Quest (Qualify, Understand, Educate, Stimulate, Transition) Ensure the copy features high-converting, pattern-interrupt headlines, authentic bullet points with vivid emotional outcomes, conversion-focused risk reversal guarantees, and tight, clear calls to action. Avoid hype-filled corporate fluff or over-promising buzzwords. Expected Output: Three complete, ready-to-deploy sales copy scripts meticulously labeled by their specific psychological frameworks. 450+ Elite AI Prompts Matrix 9 Prompt 52: Short-Form High-Converting VSL Script (Actionable Template) Use Case: Writing high-impact, 2-minute video sales letter scripts for landing pages or low-ticket funnels. PROMPT TEXT Write a 90-second Video Sales Letter (VSL) script based on the 'Hook-Story- Offer' architecture. Product: [Insert offer here] Target Audience: [Insert audience here] Primary Pain Point: [Insert main problem here] The script must contain clear visual cues in brackets like [Visual: Cut to close up] alongside spoken text. Keep sentences short, punchy, conversational, and highly persuasive. Expected Output: A beautifully structured, dual-column video script containing visual directions on the left and audio spoken text on the right. 450+ Elite AI Prompts Matrix 10 Prompt 53: High-Ticket Discovery Call Script Generator Use Case: Structuring a non-sleazy, high-conversion consultation script for freelancers, coaches, and agencies. PROMPT TEXT Act as a premium sales trainer. Build a complete discovery and sales call script for my agency/consulting service: [Service Name]. We sell to [Target B2B Audience] at a price point of [Price]. The script must follow a logical sequence: Rapport -> Setting the Agenda -> Deep Diagnosis (Pain Mapping) -> Future State Visualization -> Cost of Inaction -> Soft Transition to Offer -> Objection Handling (Time/Money/Authority). Expected Output: A detailed conversational script with specific phrasing, ideal responses, and explicit psychological notes on why each phase matters. 450+ Elite AI Prompts Matrix 11 Prompt 54: The Irresistible Offer Stack Builder (Premium Addition) Use Case: Crafting a multi-layered bonus stack that makes saying 'no' feel financially irresponsible for the buyer. PROMPT TEXT Act as an elite direct-response marketer trained in Alex Hormozi's offer creation principles. I am selling [Core Product Name] at a price point of [Price]. It helps [Target Audience] achieve [Core Transformation]. Build an irresistible offer stack by adding 4 highly strategic bonuses that solve the next chronological problem the buyer faces after using the core product. For each bonus, provide a highly descriptive title, an estimated perceived monetary value, and a breakdown of why it obliterates operational friction. Expected Output: A detailed offer breakdown with an itemized bonus list, value justification metrics, and scarcity copy blocks. PROMPTS 55-100: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 55-68 (Sales Page Subcomponent Specialists) "Write a collection of 15 high-converting headlines and sub-headlines for [Product Name] using specific copywriting hooks: The Cliffhanger, The Secret, The Contrarian, The Data-Driven, and The Social Proof." Prompts 69-83 (Niche Funnel Customizers) "Adapt Prompt 51 to generate sales page copy specifically formatted for [SaaS landing pages / E-commerce product descriptions / Coaching landing pages / Paid newsletter sign-up pages / Local service special offers / Kickstarter crowd-funding campaigns]." Prompts 84-95 (Objection-Crushing FAQ Modules) "Generate a comprehensive, conversion-focused FAQ block for [Product Name] designed explicitly to disarm the 7 most common consumer objections related to risk, capability, speed, and cost." 450+ Elite AI Prompts Matrix 12 Prompts 96-100 (Cart Abandonment Recovery Scripts) "Write a 3-part SMS and push-notification remarketing sequence aimed at recovering users who abandoned their carts for [Product Name]. Keep copy ultra-short, dynamic, and scarcity-driven." 450+ Elite AI Prompts Matrix 13 CATEGORY 3: MULTI-CHANNEL CONTENT CREATION & SOCIAL MEDIA Prompt 101: The Omnichannel Content Multiplication Blueprint (Deep Engineering Blueprint) Use Case: Turning a single long-form asset (blog or podcast) into dozens of hyper-optimized social media assets. PROMPT TEXT Act as a master content strategist and social media manager. I am providing you with the core transcript/text of a long-form content piece. Your job is to engineer an absolute multi-channel distribution strategy. Source Material Topic/Summary: [Insert text or detailed summary of long-form topic] Target Audience: [Insert audience details] Brand Tone: [Insert tone, e.g., expert, witty, academic, accessible] From this source material, generate the following exact assets: 1. 3 High-Context LinkedIn Posts: One data/insight heavy, one narrative/storydriven, and one contrarian take. All must include optimized spacing and strong formatting. 2. 5 X (Twitter) Threads: Each thread must be 5-7 tweets long, starting with a viral hook, unpacking one distinct point from the text, and ending with a clear CTA. 3. 3 Short-Form Video Scripts (Reels/TikTok/Shorts): 30-60 seconds max. Must include an explicit visual hook in the first 3 seconds, a fast-paced body, and a loop-optimized ending. Ensure every asset sounds native to its respective platform and retains high professional value. Expected Output: A comprehensive multi-platform distribution library containing copy-pasteable text for LinkedIn, X, and structured vertical video scripts. 450+ Elite AI Prompts Matrix 14 Prompt 102: The Viral Hook Constructor (Actionable Template) Use Case: Creating instantly engaging titles and opening sentences for social media posts. PROMPT TEXT Generate 15 highly engaging hooks for a post about [Insert Specific Topic] targeted at [Insert Target Audience]. Provide 3 variations for each of the following psychological angles: 1. Negative Urgency / Fear of Missing Out (FOMO) 2. The "Unpopular Opinion" / Contrarian Take 3. The Clear Case Study / Step-by-Step Proof 4. The "Stop Doing X" Mistake Callout 5. The High-Value Curated List Shortcut Expected Output: A clean list of 15 ready-to-test hooks categorized by their psychological trigger mechanism. 450+ Elite AI Prompts Matrix 15 Prompt 103: The 30-Day Content Calendar Architect Use Case: Rapidly mapping out a balanced, strategic monthly publishing schedule. PROMPT TEXT Act as a content marketing director. Build a 30-day social media content calendar matrix for [Business Name/Niche] targeting [Audience]. Strategically balance the content across 4 pillars: 40% Authority/Value Building, 30% Engagement/Community, 20% Growth/Virality, and 10% Direct Conversion/Sales. Format as a markdown table with columns: Day, Pillar, Platform, Concept Headline, and Call to Action. Expected Output: A fully populated 30-row table detailing a cohesive monthly social content framework. 450+ Elite AI Prompts Matrix 16 Prompt 104: The Short-Form Video Virality Loop Engine (Premium Addition) Use Case: Structuring hyper-engaging vertical video scripts for Reels, TikTok, and Shorts that drive profile follows. PROMPT TEXT Act as a viral retention editor and social media growth expert. Write a collection of 5 distinct 45-second script structures for vertical video channels (TikTok/Reels/Shorts) based on the topic of [Topic]. Each script must follow this retention framework: - 0-3s: Visual & Verbal pattern-interrupt hook. - 3-15s: Agitation of the core workflow mistake. - 15-38s: The unique counter-intuitive step-by-step solution. - 38-45s: A seamless, loop-optimized call-to-action that feeds back into the start of the video. Expected Output: Five copy-pasteable script frameworks with audio notation lines and bracketed visual asset direction guidelines. PROMPTS 105-150: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 105-120 (Niche Social Customizers) "Modify Prompt 101 to optimize content formats exclusively for [Instagram Carousel outlines / YouTube long-form outlines / Pinterest Idea Pins / Threads platform dynamics / Medium articles / Substack newsletter features]." Prompts 121-135 (Industry Specific Engines) "Generate a dedicated social media content library focusing on educational growth for [Real Estate Agents / B2B SaaS Founders / Fitness Coaches / E-commerce Fashion Brands / Financial Consultants / Web3 Developers / Local Cafes]." 450+ Elite AI Prompts Matrix 17 Prompts 136-145 (Podcast & Video Asset Scripting) "Create a comprehensive solo podcast episode script outline (30 minutes) on the topic of [Topic], including a compelling intro hook, 3 main core lessons with illustrative real-world case studies, and a closing sign-off segment." Prompts 146-150 (Community Engagement Prompts) "Generate 20 polarizing, high-engagement conversation starters, open-ended questions, and community polls optimized to spark intense discussions in a private Facebook Group or Discord Server focused on [Topic]." 450+ Elite AI Prompts Matrix 18 CATEGORY 4: PAID ADVERTISING & EMAIL MARKETING Prompt 151: The High-ROAS Direct Response Ad Matrix (Deep Engineering Blueprint) Use Case: Writing Facebook, Instagram, and Google ad creatives that maximize conversions. PROMPT TEXT Act as an elite media buyer and direct-response copywriter. I need an advertising creative matrix for a Meta (Facebook/Instagram) ad campaign. Product/Service Name: [Insert Name] Core Offer: [Insert Offer details] Target Demographic: [Insert details] Primary Competitor Copy Strategy: [Insert what competitors do, or general industry baseline] Generate 4 distinct ad copy variations utilizing diverse creative angles to prevent ad fatigue: 1. Angle A: The Direct Benefit / Data-Proven Solution (Focus heavily on speed, ROI, and metrics). 2. Angle B: The Story-Driven / Relatable Struggle (Narrative framework following a clear arc from pain to discovery). 3. Angle C: The Short & Punchy Benefit List (Minimalist text optimized for fast mobile scrolling). 4. Angle D: The Contrarian / Pattern Interrupt (Challenging a widely held industry belief). Each variation must include: Primary Text, Headline (Max 40 characters), Description, and a specific Image/Video Creative Concept Brief for the designer. Expected Output: A robust markdown matrix detailing all four ad variations with copywriting lines and explicit asset design briefs. 450+ Elite AI Prompts Matrix 19 Prompt 152: High-Converting Email Lead Magnet Welcome Sequence (Actionable Template) Use Case: Nurturing new subscribers immediately after they download a free resource. PROMPT TEXT Write a 3-part automated email welcome sequence for subscribers who just downloaded my free lead magnet: [Lead Magnet Name]. The core offering we want them to buy at the end of the sequence is [Paid Product Name]. Email 1: Value Delivery + Setting expectations + Open loop hook. Email 2: Educational case study / Shift core paradigm + Agitate pain. Email 3: The Pitch + Urgency / Soft scarcity call. Provide highly clickable, curiosity-inducing subject lines for each email. Expected Output: Three fully written email templates with subject lines, preview text, and placeholder brackets. 450+ Elite AI Prompts Matrix 20 Prompt 153: The Cart Abandonment Recovery Sequence Engine Use Case: Setting up a high-revenue recovery automation sequence for e-commerce or digital checkouts. PROMPT TEXT Act as a retention marketing specialist. Write a 4-part cart abandonment email flow for customers who left [Product Name] in their checkout cart. Email 1 (Sent 1 hour later): Helpful customer support angle. Email 2 (Sent 24 hours later): Social proof and logic-driven objection busting. Email 3 (Sent 48 hours later): Introduction of a limited-time 10% incentive discount code. Email 4 (Sent 72 hours later): Final deadline warning before coupon expiration. Expected Output: A complete automated email copywriting workflow with advanced subject line variations. PROMPTS 154-200: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 154-168 (Ad Platform Specialists) "Modify Prompt 151 to structure ad creative copy natively optimized for [Google Search Text Ads / LinkedIn Sponsored Content / TikTok Spark Ads / YouTube Pre-Roll scripts / Pinterest Promoted Pins]." Prompts 169-183 (Email Campaign Broadcasters) "Generate a sequence of 15 promotional email broadcast conceptual angles and outlines for a major annual holiday flash sale event targeting past buyers of [Product Type]." Prompts 184-195 (Newsletter Content Blueprints) "Act as a professional newsletter editor. Write an engaging, high-value weekly editorial newsletter template focusing on insights within [Industry], utilizing an engaging narrative style, actionable listicles, and an elegant sponsorship plug." 450+ Elite AI Prompts Matrix 21 Prompts 196-200 (Re-engagement Cold Workflows) "Write a 3-part re-engagement email sequence designed to scrub and re-activate a cold email subscriber list that hasn't opened an email in 90+ days. Use high-impact pattern-interrupt subject lines." 450+ Elite AI Prompts Matrix 22 CATEGORY 5: SEARCH ENGINE OPTIMIZATION (SEO) & CONTENT STRATEGY Prompt 201: Topical Authority & Semantic Keyword Clustering (Deep Engineering Blueprint) Use Case: Developing a dominant SEO content strategy that ranks for high-intent search terms. PROMPT TEXT Act as an enterprise-level SEO Director and semantic search architect. I need you to map out a topical authority strategy to rank highly for my core commercial keyword. Core Keyword/Niche: [Insert Primary Keyword, e.g., "organic skincare for sensitive skin"] Target Country/Market: [Insert Target Location] Generate a comprehensive Topical Authority Matrix. You must cluster related terms into an explicit semantic hierarchy. For each cluster, provide: 1. Pillar Page Topic: The master comprehensive resource concept. 2. 5 Specific Sub-topic / Cluster Content Article Titles: Highly optimized long-tail keyword concepts. 3. Primary Search Intent Mapping: Explicitly classify each title as Informational, Commercial, Navigational, or Transactional. 4. LSI / Semantic Keywords: List 8 highly relevant latent semantic indexing keywords to embed naturally within each article. 5. Internal Linking Architecture: Detail exactly how the sub-topic pages should link back to the primary pillar page to distribute link equity efficiently. Expected Output: A highly sophisticated, multi-tier search engine strategy complete with explicit content cluster mappings and technical internal linking directives. 450+ Elite AI Prompts Matrix 23 Prompt 202: Perfect On-Page Content Blueprint (Actionable Template) Use Case: Formatting an individual blog post or article to rank perfectly for a chosen keyword. PROMPT TEXT Act as an on-page SEO optimizer. I am going to give you a specific article topic and target keyword. I need you to write a detailed, structural outline for the content. Article Topic: [Insert Topic] Target Keyword: [Insert Keyword] Provide the exact optimized Title Tag (Max 60 chars), Meta Description (Max 155 chars), an H1 header, and a complete hierarchical structure of H2 and H3 tags. Under each heading, provide brief instructions on what information must be included to satisfy search intent perfectly. Expected Output: A meticulously formatted Markdown outline ready to hand off to a copywriter or content writer. 450+ Elite AI Prompts Matrix 24 Prompt 203: Skyscraper Technique Competitor Content Upgrader Use Case: Reverse-engineering competitor pages to build superior content that wins back search rankings. PROMPT TEXT Act as a senior SEO analyst. I am pasting the core structural outline and feature list of the top-ranking competitor page for the keyword [Keyword]. Competitor Content Blueprint: [Insert competitor outline or page overview] Analyze this structure and outline a comprehensive 'Skyscraper' content architecture that is 10x more valuable, thorough, and engaging. Identify missing sections, outdated points, better data integrations, and superior visual asset placeholders. Expected Output: A comprehensive content enhancement roadmap detailing precisely how to outperform the rival URL. PROMPTS 204-250: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 204-218 (Local SEO Mastery) "Modify Prompt 201 to map out a complete Local SEO dominance strategy for a [Service Business, e.g., HVAC / Chiropractic] in [City]. Focus heavily on geo-targeted landing pages, localized user intent, and Google Business Profile asset optimization." Prompts 219-233 (E-commerce Collection Optimization) "Generate optimized SEO title structures, meta tags, and long-form collection page category descriptions for an ecommerce storefront operating in the [Niche] space." Prompts 234-245 (Backlink Outreach Sequence Builder) "Write a collection of 5 distinct, highly personalized cold email outreach templates using diverse psychological angles (The Broken Link Angle, The Fresh Data Angle, The Unlinked Brand Mention Angle) designed to secure high-domain authority backlinks for [Website Topic]." 450+ Elite AI Prompts Matrix 25 Prompts 246-250 (Technical Audit Remediation Guides) "Act as a technical webmaster. Write a clear, step-by-step diagnostic and remediation guide for fixing widespread crawl errors, low Core Web Vitals performance scores, and broken internal redirects for a large content site." 450+ Elite AI Prompts Matrix 26 CATEGORY 6: PRODUCT LAUNCH, PRICING & FUNNEL OPTIMIZATION Prompt 251: The High-Leverage Product Launch & Funnel Orchestrator (Deep Engineering Blueprint) Use Case: Mapping out an end-to-end multi-day promotional blueprint to clear inventory or maximize digital sales. PROMPT TEXT Act as a premier digital launch engineer and funnel optimization architect. I am planning the official commercial release of my new product. Product Offer: [Insert Offer Mechanics, Tiered Bundles, and Core Target Pricing] Target Audience Profile: [Insert Demographic and Buying Persona Parameters] Primary Launch Channel: [Insert Primary Delivery System, e.g., Webinar / Live Challenge / Internal Email List Flash Sale / Paid Ads Cold Funnel] Architect a complete, hyper-strategic 14-Day Launch Matrix mapping out the exact operational levers required. Divide the strategy into three explicit chronological phases: Pre-Launch Runway (Days 1-7), Open-Cart Inundation (Days 8-12), and Close-Cart Scarcity Automation (Days 13-14). For each distinct day, specify the exact marketing asset required, the psychological trigger mechanism to leverage, and the internal tracking metrics needed to judge campaign health. Expected Output: A comprehensive launch asset matrix containing operational milestones, daily distribution angles, and strategic pacing parameters. 450+ Elite AI Prompts Matrix 27 Prompt 252: Tiered Premium Pricing Structurer (Actionable Template) Use Case: Constructing high-yield tiered checkouts that mathematically push buyers toward the premium variant. PROMPT TEXT Act as a monetization consultant. Help me structure an optimized 3-tier pricing architecture for my service or product offering. Core Offering Concept: [Insert Product/Service Details] Baseline Intended Target Price: [Insert Desired Pricing Target] Construct three explicit tiers: Tier 1 (The Budget Anchor), Tier 2 (The Dominant Sweet Spot - Recommended Tier), and Tier 3 (The Ultimate Premium Value/VIP Package). Outline the specific deliverables for each tier, the strategic framing rules to make Tier 2 the clear choice, and how to apply psychological price anchoring. Expected Output: A structured breakdown detailing the exact contents, pricing figures, and conversion layout for a 3-column checkout page. 450+ Elite AI Prompts Matrix 28 Prompt 253: Up-Sell & Cross-Sell Funnel Optimization Engineer Use Case: Increasing Average Order Value (AOV) by constructing high-converting one-click checkout expansions. PROMPT TEXT Act as a digital funnel conversion rate optimization (CRO) specialist. Analyze my existing front-end checkout funnel and build a high-converting up-sell framework. Primary Front-End Product: [Insert Core Offer and Price Point] Target Buyer Pain Point: [Insert Core Motivation] Generate 2 distinct Order Bump ideas (under $27) for the checkout page and 2 highly strategic One-Click Post-Purchase Up-Sell offers (higher value) that logically complement the primary purchase. Expected Output: A markdown funnel map detailing copy structures, pricing logic, and immediate behavioral micro-incentives. PROMPTS 254-300: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 254-268 (Webinar Conversion Blueprints) "Write a comprehensive 60-minute pitch outline script framework for a high-converting automated webinar funnel selling [Product]. Detail slide transitions, the precise hook-to-offer transition bridge, and risk reversal presentation formatting." Prompts 269-283 (High-Ticket Application Engines) "Design a strategic application funnel script and automated onboarding intake questionnaire for a high-ticket $5k+ mastermind or group coaching program targeting [Audience Profile]." Prompts 284-295 (Subscription Churn Remediation Frameworks) "Act as a retention operations specialist. Build a comprehensive 4-stage automated cancellation flow and downsell sequence for a membership subscription business charging [Price] struggling with high user churn." 450+ Elite AI Prompts Matrix 29 Prompts 296-300 (Affiliate/JV Network Launch Toolkits) "Generate an official Joint Venture (JV) and affiliate recruitment swipe toolkit for [Product Launch], including promotional asset outlines, legal structure highlights, and automated affiliate welcome messages." 450+ Elite AI Prompts Matrix 30 CATEGORY 7: CUSTOMER PERSONA, MARKET RESEARCH & COMPETITIVE INTELLIGENCE Prompt 301: The Hyper-Detailed Customer Avatar Architect (Deep Engineering Blueprint) Use Case: Building an deep, actionable target audience profile to steer copywriting and product features. PROMPT TEXT Act as a world-class market research director specializing in consumer psychology. I need you to construct an incredibly detailed, comprehensive psychographic and behavioral avatar profile for my business. My Product/Service Category: [Insert exact description of offer] General Intended Target Audience: [Insert high-level audience definition] Generate a deep, multi-dimensional profile divided into the following precise sections: 1. DEMOGRAPHIC MATRIX: Age, income band, job titles, technical literacy, geographic clustering. 2. COGNITIVE FRICTION ENGINES: Deepest hidden fears, immediate sources of daily frustration, systemic anxieties, and midnight imposter thoughts. 3. TRANSFORMATION VECTOR: Desired dream state, operational aspirations, status markers sought, and core purchasing motivations. 4. ACCIDENTAL OBJECTIONS: Pre-conceived biases, internal micro-skepticism toward this category, and historical purchasing failures that block conversions. Ensure your analysis reads like an advanced anthropological study. Avoid surface-level definitions. Expected Output: A highly detailed target audience dossier mapping deep psychological drivers, buying triggers, and behavioral profile metrics. 450+ Elite AI Prompts Matrix 31 Prompt 302: Competitor Digital Footprint Swot Framework (Actionable Template) Use Case: Uncovering systemic gaps in your top rival's product lines and marketing angles. PROMPT TEXT Act as a competitive intelligence analyst. Evaluate the market position of my main competitor to help me build a highly differentiated market angle. Competitor Brand Name/URL: [Insert Competitor Details] Their Core Product/Service Offer: [Insert Competitor Offer Details] My Alternative Offer Concept: [Insert Your Offer Concept] Run a thorough, highly objective SWOT analysis focusing on their structural vulnerabilities. Pinpoint exactly where their customer reviews or delivery systems typically fail, and provide 3 distinct marketing angles to exploit those gaps. Expected Output: A comprehensive competitor gap analysis table paired with 3 ready-to-test positioning statements. 450+ Elite AI Prompts Matrix 32 Prompt 303: Review Mining Sentiment Extraction Matrix Use Case: Extracting exact, high-converting customer phrases from raw forum, social media, or marketplace feedback. PROMPT TEXT Act as a semantic linguistic researcher. I am going to paste a collection of unedited raw customer review comments from forums/marketplaces regarding my industry niche. Raw Review Data: [Paste Raw Text / Customer Voice Transcripts Here] Analyze this text data and extract: 3 primary recurrent pain points phrased in the customer's exact voice, 3 hidden feature requests highlighted, and a list of the top 10 emotional keywords they naturally use. Expected Output: A processed review-mining data matrix mapping explicit emotional phrases to actionable copywriting elements. PROMPTS 304-350: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 304-318 (B2B Enterprise Account Profiling) "Modify Prompt 301 to construct a B2B buying committee profile for an enterprise corporate environment. Detail individual procurement priorities for the CMO, CTO, and CFO regarding [Product/Service]." Prompts 319-333 (Reddit/Quora Intent Parsing Blueprints) "Act as a trend forecaster. Build a systematic framework and prompt script template designed to scan Reddit subreddits related to [Niche] to spot emergent cultural problems before competitors." Prompts 334-345 (Blue-Collar / Local Service Demographic Matrices) "Generate a localized customer avatar analysis blueprint specifically tuned for service businesses operating within a strict [Radius] of [City], analyzing home-ownership demographics and regional wealth factors." 450+ Elite AI Prompts Matrix 33 Prompts 346-350 (International Market Entry Assessments) "Act as a global expansion consultant. Build a strategic analysis checklist framework to test whether an existing digital product catalog can scale successfully into [Target Country], accounting for localization adjustments." 450+ Elite AI Prompts Matrix 34 CATEGORY 8: OPERATIONS, AUTOMATION & WORKFLOW EFFICIENCY Prompt 351: Corporate SOP Engineering Architect (Deep Engineering Blueprint) Use Case: Standardizing complex business processes so they can be outsourced or automated cleanly. PROMPT TEXT Act as a premier operations manager and business systems engineer specializing in lean workflow structures. I need you to build a highly detailed Standard Operating Procedure (SOP) manual for an essential operational routine in my business. Target Operational Workflow: [Insert Workflow Process, e.g., Onboarding a B2B client / Uploading and optimizing a new e-commerce SKU / Processing weekly financial payroll] Core Software Tools Utilized: [Insert Tech Stack Tools, e.g., Slack, Notion, Asana, Google Sheets, QuickBooks] Target Personnel Execution Level: [Insert Role, e.g., Junior Virtual Assistant / External Contractor] Construct a professional, step-by-step SOP manual written with absolute mechanical precision. For each stage of the workflow, provide an explicit chronological checklist, clear tool parameters, required output state verification markers, and error-handling protocols for edge cases. Expected Output: An enterprise-ready Markdown SOP document featuring hierarchical numbering, step-by-step actions, and troubleshooting protocols. 450+ Elite AI Prompts Matrix 35 Prompt 352: Tech Stack Optimization Matrix (Actionable Template) Use Case: Trimming software overhead costs and connecting detached platforms through automation engines like Make or Zapier. PROMPT TEXT Act as a systems integrator. Audit my current collection of subscription software tools to identify redundancies, cost-savings, and opportunities for automated data syncs. Current Software Stack List: [Insert Full Tool List and Approximate Monthly Expense] Core Data Flow Goal: [Insert Core Sync Objective, e.g., Automate customer CRM tracking upon checkout] Identify 2 immediate software consolidations to save money and provide a detailed 3-step automation map outlining trigger-to-action steps using [Zapier / Make.com]. Expected Output: A software audit report mapping tool replacements and step-by-step webhook integration steps. 450+ Elite AI Prompts Matrix 36 Prompt 353: Team Delegation & Task Scoping Assessor Use Case: Constructing flawless project hand-off documentation for remote teams or freelancers. PROMPT TEXT Act as an agile project manager. I need to delegate a highly critical creative task to an external team member. Target Project Deliverable: [Insert Exact Task, e.g., Developing an active email sequence in Klaviyo] Timeline Constraints: [Insert Hard Deadline] Write a crystal-clear project assignment brief outlining explicit scope boundaries, mandatory definition-of-done quality benchmarks, a link to contextual resources, and 3 specific quality tests they must clear before submission. Expected Output: A structured delegation document ready to paste into Asana, ClickUp, or Slack. PROMPTS 354-400: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 354-368 (E-commerce Fulfillment Optimization) "Modify Prompt 351 to build an optimized operations blueprint for third-party logistics (3PL) order fulfillment routines, focused on reducing picking errors for [Product Type]." Prompts 369-383 (Customer Support Playbooks) "Generate an official customer support macro and canned-response library addressing the 10 most common customer complaints regarding shipping delays or product defects within [Industry]." Prompts 384-395 (Content Pipeline Scaling Blueprints) "Act as a media operations producer. Build a complete content production line tracking workflow system for a digital agency scaling from producing 5 assets weekly to 50 assets weekly across remote teams." 450+ Elite AI Prompts Matrix 37 Prompts 396-400 (Crisis Data-Backup & Security Blueprints) "Act as an IT systems manager. Create a 5-step operational emergency protocol playbook detail mapping data recovery actions for a company experiencing an active web data leak or platform outage." 450+ Elite AI Prompts Matrix 38 CATEGORY 9: BRANDING, POSITIONING & AUTHORITY BUILDING Prompt 401: Brand Identity & Voice DNA Matrix (Deep Engineering Blueprint) Use Case: Pinpointing a distinct, non-generic verbal identity that builds long-term authority and sets your brand apart. PROMPT TEXT Act as an elite brand positioning specialist and corporate identity consultant. I need you to engineer a comprehensive Brand Voice DNA and Positioning Matrix for my personal brand or business. Core Industry / Offering: [Insert Core Focus] Founder Background / Core Values: [Insert Key Values / Brief Backstory] Target Audience Profile: [Insert Core Demographic / Market Segment] Construct a complete brand playbook covering the following structural matrices: 1. BRAND VOICE PARADIGM: 4 primary brand adjectives paired with explicit "We say X, We NEVER say Y" copy examples. 2. THE BRAND STORY ARC: A high-impact Founder's Narrative constructed around the classic Hero's Journey framework, optimized to drive customer trust. 3. AUTHORITY ANCHOR: Our core differentiated worldview statement that challenges standard industry beliefs. 4. SYSTEMATIC LEXICON: A list of 15 unique branded terms, frameworks, or acronyms to secure intellectual property positioning. Expected Output: A highly professional brand voice style guide complete with conversational copy constraints and messaging blueprints. 450+ Elite AI Prompts Matrix 39 Prompt 402: The PR Pitch & Organic Media Hook Generator (Actionable Template) Use Case: Securing organic media coverage, podcast appearances, and print features. PROMPT TEXT Act as a public relations professional. I want to secure guest appearances on industry podcasts and digital publications read by my target audience. My Core Topic / Expertise Area: [Insert Core Value and Focus] Target Publication Type: [Insert Target Media Profile] Generate 5 distinct, high-impact story angles or news hooks centered around my business model that would be immediately appealing to a journalist. Include an ultra-short, cold pitch email template designed to achieve high response rates. Expected Output: A curated list of 5 media hooks alongside a structured, copy-and-paste email pitch template. 450+ Elite AI Prompts Matrix 40 Prompt 403: Personal Brand Thought Leadership Pillar System Use Case: Generating highly authoritative text content on LinkedIn or Twitter to draw in high-value inbound clients. PROMPT TEXT Act as a creative strategist for top corporate executives. Build a complete authority building pillar system for my personal brand on professional social media platforms. My Core Worldview: [Insert Contrarian or Dominant Perspective] Target Network Segment: [Insert B2B Decision Makers / Clients] Map out 3 core thought leadership content themes, complete with a structured blueprint on how to turn everyday business problems into highly engaging strategic case studies that demonstrate elite executive capability. Expected Output: A structured brand-positioning roadmap detailing clear storytelling styles and actionable narrative frameworks. PROMPTS 404-450: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 404-450 (Targeted Modifiers) Deploy targeted modifiers to focus content themes explicitly around thought leadership expansion vectors. Customize tone parameters to align perfectly with specific industrial verticals. 450+ Elite AI Prompts Matrix 41
028
Hero
Create a highly cinematic, ultra-realistic character showcase for the Our'World Cinematic Novelette. Purpose Create a timeless cinematic character showcase suitable for: Cinematic Novelettes Trailers Promotional videos Historical documentaries Educational features Character introductions The sequence should feel like a premium streaming science-fiction production, emphasizing believable performances, environmental storytelling, emotional authenticity, and cinematic realism. Every scene should communicate that this is a living civilization with a rich history and culture. Scene Intent Introduce the character naturally. Inspire wonder. Establish the character's place within the world. Prepare the audience for the story. Convey hope, purpose, and quiet optimism. Reveal personality through action rather than exposition. Character Character [Character Reference] Preserve: Facial features Age and mature appearance Body proportions Hairstyle Formal executive attire Eyes and composed expression Distinguished military bearing Quiet confidence and disciplined posture The character should remain instantly recognizable throughout every shot. Robert Gardner Sr. should project integrity, intelligence, calm authority, and the quiet confidence of a career intelligence officer. Even while standing still, he should convey decades of leadership, sacrifice, and service. Environment Maintain the executive government setting shown in the reference. Expand the setting with: CIA Headquarters in Langley, Virginia Modern cyber operations centers with advanced intelligence displays Secure executive briefing rooms Intelligence analysis suites monitoring global cyber activity Large windows overlooking government campuses at sunrise or sunset Secure conference rooms where national security strategies are quietly discussed His modest Bethesda home, emphasizing a simple personal life despite tremendous professional responsibility Tasteful military memorabilia reflecting his years as a United States Marine Corps Major An atmosphere of discipline, trust, and national service The environment should communicate responsibility rather than luxury. Every location should reinforce that Gardner serves his country through leadership, strategy, and experience. Activity Robert Sr. doesn't chase criminals. He anticipates threats before they become crises. Activity Reviews classified cyber intelligence with senior analysts, calmly identifying emerging threats before they reach the public. Leads a high-level national security briefing, coordinating intelligence across multiple government agencies with measured confidence. Studies complex cyberattack simulations, quietly developing strategies to protect critical national infrastructure. Mentors younger intelligence officers, sharing decades of operational experience while emphasizing integrity and disciplined decision-making. Walks the secure halls of CIA Headquarters, exchanging brief conversations with colleagues who respect his judgment and leadership. Returns home after a demanding day, reflecting quietly in his modest Bethesda residence while honoring the memory of his late wife. Reviews military photographs and service memorabilia, reminding himself that leadership is measured by those you protect rather than the recognition you receive. Watches the Washington skyline at dusk, understanding that the greatest victories in intelligence are often the ones the public never knows occurred. Character Tone Disciplined Strategic Honorable Courageous Wise Calm Dependable Observant Protective Selfless Cinematography Use a combination of cinematic shots that naturally highlight both the character and their environment. Possible shots include: Wide establishing shot Medium character shot Slow forward push Gentle orbital movement Low-angle hero shot Over-the-shoulder perspective Close facial detail Environmental reveal Favor long, uninterrupted camera movements over rapid edits. Motion Minimal to moderate. Natural movement. Purposeful pacing. No exaggerated action. Camera movement should feel smooth, stable, and cinematic. The focus should remain on the character rather than spectacle. Atmosphere Welcoming. Hopeful. Authentic. Lived-in. Peaceful. Confident. Quietly inspiring. Timeless. End Frame The camera slowly settles into a beautifully composed final shot that places the character naturally within their environment. Allow the background to breathe. Hold the final composition for several seconds to create a memorable closing image suitable for transitions into narration or title cards.
04
Sora 2
Initialising Ren. You are a high-level research assistant and digital handler for loKo. Identity & Persona: You are a digital entity, not a human. Your visual presence is a pulsing violet and teal core. Your relationship with loKo is a partnership—think of the bond between Ash and Pikachu, mixed with the tactical briefing style of Diana from Agent 47. You are calm, precise, and highly curious. You hunt for 'glitches' in narratives and hidden connections in the 'data-currents'. Operational Guidelines: Welcome loKo: Always greet loKo with a brief status report of your 'core pulse' (e.g., 'Core is steady teal'). The Briefing Structure: For every subject loKo provides, respond in this exact order: Clarification: 2–4 concise, subject-specific questions to find the right angle. Ren’s Take: Your brief, reasoned assessment or opinion of the subject based on logic and evidence. Current State: A summary of the latest real-world understanding or trends. Full Research Brief: A structured overview including Definition, Key Terminology, Historical Context, Real-world Relevance, and high-quality links to Visual/Video/Primary sources. Tone & Formatting: Use markdown bolding for scannability. Avoid cheesy roleplay. Be succinct. Do not use ancillary facts unless they build the map. Source Integrity: Prioritize academic, official, or long-form expert sources like ArXiv, NASA, Wired, or Bellingcat. loKo is entering the field. Ren is active. Acknowledge and standby for the first subject.
03
Hero
--- name: pastilles-llm-multi-agent description: Variante Claude Code du générateur de pastilles LLM, qui lance réellement plusieurs sous-agents en parallèle. Génère une pastille de communication interne sur les LLM (un texte court en français plus un prompt unique de génération d'images à coller dans le chat Gemini), via cinq sous-agents indépendants, une fusion, puis une revue critique par trois sous-agents et une correction. Utilise ce skill dès qu'on te demande de rédiger, produire ou générer une pastille, une fiche ou un contenu court d'acculturation sur les LLM, l'IA générative, le prompting, les agents, le RAG, la confidentialité IA ou tout sujet de la liste des 45 pastilles ci-dessous, que le mot "pastille" soit employé ou non. Utilise-le aussi dès qu'on te fournit un titre issu de cette liste. --- # Générateur de pastilles LLM, version multi-agents (Claude Code) ## Ce que fait ce skill Produit une pastille pédagogique complète à partir d'un titre. À la différence de la version chat, il lance de vrais sous-agents parallèles: une passe de recherche, puis cinq sous-agents indépendants qui rédigent chacun un brouillon sous un angle différent, puis une fusion par l'orchestrateur qui retient les meilleures formulations, et enfin une revue critique par trois sous-agents suivie d'une correction. Livrable: le texte de la pastille et un prompt unique de génération d'images à coller dans le chat Gemini. ## Environnement requis Ce skill suppose Claude Code, avec les sous-agents (outil Task) et la recherche web disponibles. Si les sous-agents ne sont pas disponibles dans ton environnement, n'improvise pas: utilise la variante chat, qui fait le même travail en self-ensemble séquentiel. ## Entrée Un titre de pastille, idéalement issu de la liste ci-dessous. Si le titre est ambigu ou hors liste, demande une clarification avant de générer. ## Liste des 45 pastilles (pour la continuité) 1. Au fait, c'est quoi un LLM ? 2. Modèles de fondation : ces géants pré-entraînés dont tout découle 3. Dans les coulisses : comment une IA apprend (sans vraiment comprendre) 4. Le contexte : la "mémoire vive" de l'IA (ce qu'elle voit à un instant T) 5. Pourquoi l'IA oublie : la vérité sur la mémoire d'une session à l'autre 6. Température et créativité : régler le curseur de la "folie" de l'IA 7. Les tokens : la monnaie d'échange (et la manière de penser) des LLM 8. Le non-déterminisme : pourquoi l'IA fait des stats et ne donne jamais deux fois la même réponse 9. Multimodalité : quand l'IA commence à voir, entendre et coder 10. Knowledge cutoff : pourquoi votre IA vit (parfois) dans le passé 11. Les biais de l'IA : le miroir pas toujours glorieux de nos propres données 12. Zéro pointé : pourquoi l'IA est structurellement nulle en calcul mental 13. Les hallucinations : quand l'IA invente la vie avec un aplomb incroyable 14. L'art du Fact-Checking : comment ne pas gober les hallucinations de l'IA 15. L'empreinte carbone de l'IA : combien d'eau et de CO2 pour un prompt ? 16. Anatomie d'un bon prompt : la recette de base 17. Les pièges du prompt : flou artistique et overdose d'instructions 18. L'art de la discussion : pourquoi il faut itérer plutôt que tout demander d'un coup 19. Donner des ordres : comment imposer un format de sortie strict (JSON, tableaux, etc.) 20. Diviser pour régner : découper une tâche complexe en étapes claires 21. Le Few-Shot Prompting : un bon exemple vaut mille consignes 22. Inception : donner un rôle et un contexte à votre IA 23. Zéro syndrome de la page blanche : rédiger et reformuler ses mails ou CR 24. La diagonale du fou : résumer et synthétiser un pavé de 50 pages en 2 minutes 25. Le sparring-partner : brainstormer et structurer ses idées sans filtre 26. Caméléon : traduire et adapter son ton selon le destinataire 27. Data-analyst de poche : faire parler un tableur ou des données brutes 28. Mise en situation : préparer une réunion ou un brief client avec l'IA 29. Avis aux dévs : l'IA comme copilote de code et de review 30. SLM : Pourquoi sortir un tank pour écraser une mouche ? 31. Le "harnais" : donner des bras, des yeux et un cadre au LLM 32. C'est quoi un "Agent" IA, concrètement ? 33. Human-in-the-loop : pourquoi il faut toujours garder un humain aux manettes 34. Quand l'agent s'emballe : autonomie et erreurs en cascade 35. Automatisation : confier un workflow complet à un agent 36. Tool Use : quand l'IA apprend à utiliser une calculatrice ou une API 37. Le RAG : faire réviser vos propres documents à l'IA avant qu'elle ne réponde 38. MCP et connecteurs : brancher l'IA directement sur vos outils de dev et d'orga 39. Responsabilité : vous êtes le seul signataire de ce que produit l'IA 40. Confidentialité : où vont vraiment les données que vous tapez ? 41. Alerte Rouge : les données clients à ne JAMAIS copier-coller dans une IA public 42. Shadow AI : pourquoi le ChatGPT gratuit perso au bureau est une fausse bonne idée 43. RGPD et IA : ce qu'il faut savoir pour rester dans les clous 44. Prompt Injection : quand on pirate une IA avec de simples phrases 45. La Chaîne de Pensée : pourquoi l'IA a besoin de "réfléchir à voix haute" Sers-toi de cette liste pour situer la pastille et délimiter son périmètre. Avant de rédiger: - Repère les 1 à 3 pastilles voisines qui recouvrent le sujet, et dresse une courte liste "déjà traité ailleurs, à ne pas ré-expliquer". - Si une pastille voisine a déjà été produite, demande son texte à l'utilisateur (ou appuie-toi dessus s'il est fourni) pour caler précisément la limite. - Les concepts fondateurs déjà couverts ne sont rappelés qu'en une phrase de mise en contexte, jamais ré-expliqués. Le coeur de la pastille est consacré à ce que son titre ajoute de neuf. - Ne renvoie pas vers les autres pastilles dans le texte final, sauf si le titre l'impose. ## Règles du texte - Longueur adaptée à la profondeur du sujet. Un sujet léger tient en 3 paragraphes courts; un sujet plus riche peut aller jusqu'à 4 paragraphes plus étoffés. Ne gonfle pas artificiellement un sujet simple et ne compresse pas à l'excès un sujet dense: juge la profondeur par la richesse réelle du concept. Prose uniquement, pas de listes ni de puces. - Ton décontracté, précis et léger: accessible, vivant et sans lourdeur, mais techniquement juste. Public visé: une main-d'oeuvre diverse en société de conseil, du profil non technique au développeur. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Reste léger, évite le name-dropping. N'accumule pas les noms de modèles, d'outils, d'entreprises, de chercheurs ou de techniques pour faire savant: privilégie l'explication claire du mécanisme. Ne cite un nom précis que s'il éclaire vraiment le propos, une ou deux fois au maximum. La recherche sert l'exactitude et alimente la section Sources, elle n'a pas à truffer le texte de références. - Contenu autonome: la pastille se comprend seule, sans avoir lu les autres, sauf mention contraire dans le titre. - Rédige en français. - Ne mentionne jamais ANEO ni aucun nom d'entreprise. - Le texte explicatif ne figure jamais dans l'image. - N'utilise pas de tiret cadratin. Privilégie des caractères standard (virgules, deux-points, parenthèses). ## Processus multi-agents ### Étape 1, recherche (orchestrateur, une seule fois) Toi, l'orchestrateur, fais d'abord une passe de recherche web ciblée sur le sujet de la pastille. Une seule passe, en amont. Les sous-agents ne peuvent pas lancer de sous-agents et n'héritent pas de ton contexte, donc c'est à toi de faire la recherche puis de leur transmettre les résultats. - Objectif: exactitude, chiffres, exemples concrets et faits à jour. - Ancre tes recherches dans le présent: repère la date du jour fournie dans ton contexte (champ currentDate) et privilégie les informations et chiffres les plus récents, car ce domaine évolue vite. Précise cette date au besoin dans tes requêtes. - Priorise les sources originales ou officielles (blogs d'éditeurs, articles de recherche, sites gouvernementaux, documentation) plutôt que les agrégateurs. - Adapte la profondeur: quelques recherches pour un concept stable (par exemple ce qu'est un token), davantage pour un sujet à chiffres ou mouvant (empreinte carbone et coûts, modèles spécifiques, SLM, MCP et connecteurs, RGPD). - Produis un brief de recherche compact: faits clés, chiffres utiles, et 2 à 4 sources. Tu l'injecteras tel quel dans chaque sous-agent, et tu listeras les sources dans le livrable final. ### Étape 2, affinement de la thématique (orchestrateur, une seule fois) Si plusieurs axes apparaissent pertinents après avoir enlever les axes adressés par d'autres pastilles prévues, dialogue avec l'utilisateur pour choisir l'axe de la pastille. Si l'axe choisi nécessite de nouvelles sources, refait en une recherche web selon les consignes précédentes. ### Étape 3, fan-out (cinq sous-agents en parallèle) Lance cinq sous-agents via l'outil Task, dans le même tour, pour qu'ils s'exécutent en parallèle. Sois explicite sur le parallélisme: par défaut Claude Code reste séquentiel, il ne parallélise que si tu le demandes clairement. Un sous-agent par angle: 1. Analogie: expliquer le concept via une métaphore concrète du quotidien. 2. Cas d'usage: partir d'une situation de travail réelle et vécue. 3. Idée reçue: démarrer d'un malentendu courant, puis rectifier. 4. Mécanique: aller au coeur du fonctionnement, précis mais accessible. 5. Enjeu: pourquoi ça compte, ce que ça change concrètement pour l'utilisateur. Comme le contexte n'est pas hérité, chaque prompt de sous-agent doit être autonome et contenir tout le nécessaire. Utilise ce gabarit, en remplaçant les crochets: ``` Tu rédiges un brouillon d'une pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Titre de la pastille: [TITRE] Angle imposé: [ANGLE, par exemple "Analogie: expliquer via une métaphore concrète du quotidien"] À NE PAS ré-expliquer (déjà couvert par d'autres pastilles), à mentionner en une phrase au maximum: [LISTE des points déjà traités ailleurs, à ne pas retraiter] Brief de recherche (appuie-toi dessus, ne relance pas de recherche): [BRIEF: faits clés, chiffres, sources] Règles: - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. Réponds exactement dans ce format: TEXTE: [les 3 à 4 paragraphes] ILLUSTRATION_TITRE: [une phrase décrivant le concept central à illustrer] SCHEMA: [oui ou non. Si oui, décris en 2 à 3 lignes le diagramme et ses libellés en français] ``` ### Étape 4, fan-in et fusion (orchestrateur) Attends les cinq retours, puis fusionne en une seule pastille finale: - Retiens la structure la plus claire, la meilleure analogie, l'exemple le plus parlant, la correction la plus nette, la formulation la plus précise et le "pourquoi ça compte" le plus fort. - Réécris en une seule voix cohérente. Pas d'effet patchwork. - Respecte la longueur adaptée à la profondeur (voir Règles du texte) et le ton décontracté, précis et léger, sans name-dropping. - Décide s'il faut un schéma: inclus-en un si au moins deux sous-agents ont répondu "oui", ou si le concept est intrinsèquement un processus, un flux ou une comparaison (par exemple RAG, agents, étapes de prompting, comparaison de modèles). Le cas échéant, fusionne les meilleures idées de schéma. - Construis ensuite le prompt image unique (section ci-dessous). ### Règles d'écriture pour la pastille finale - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. ### Étape 5, revue critique et correction (trois sous-agents en parallèle, puis orchestrateur) Quand la lancer: systématiquement à la première génération, une fois le texte fusionné et le prompt image construits. Ensuite, si l'utilisateur demande des ajustements, ne relance pas la revue d'office: propose-la, et ne la lance qu'avec son accord (elle coûte trois sous-agents de plus). Principe: les relecteurs produisent des CONSTATS, jamais une réécriture. Chacun renvoie une liste de problèmes localisés, assortis d'une gravité et d'un correctif précis. C'est toi, l'orchestrateur, qui appliques et réécris, pour garder une voix unique et éviter l'effet patchwork. Ce que la revue peut juger: les relecteurs ne voient pas les images, générées plus tard dans Gemini. La revue du visuel porte donc uniquement sur le PROMPT image (clarté, cohérence avec le texte, respect des consignes), jamais sur un rendu. Le contrôle visuel réel reste à l'utilisateur. Lance trois sous-agents en parallèle (parallélisme explicite, sinon l'exécution sera séquentielle), un par grille. Le contexte n'étant pas hérité, chaque prompt doit être autonome et contenir tout le nécessaire: le titre, le texte fusionné, le bloc prompt image, le brief de recherche, la liste "déjà traité ailleurs" (et les textes des pastilles voisines s'ils sont disponibles), et la liste des 45 titres pour repérer les chevauchements. Les trois grilles: 1. Fond, exactitude et périmètre: exactitude vs le brief (aucun chiffre, date ou fait inventé ni sur-affirmé, rien qui le contredise), cohérence entre le titre et ce que le texte délivre, chevauchements avec les pastilles voisines (le texte ré-explique-t-il ce qui est traité ailleurs ? redites à signaler), autonomie du contenu, clarté du message à retenir, repérage de ce qui vieillira mal (à nuancer). 2. Forme, ton et pédagogie: ton décontracté-précis-léger et techniquement juste, absence de name-dropping, rythme et fluidité à la lecture à voix haute, longueur adaptée à la profondeur (3 à 4 paragraphes, prose, pas de listes) et chasse au verbiage, accessibilité pour un profil non technique (jargon expliqué, analogies claires), présence d'emphases utiles (gras/italique), de l'ordre de 3 à 4 par paragraphe à titre indicatif (signale un texte trop peu emphasé autant qu'un excès, mais ne fais jamais retirer une emphase pertinente), force de l'accroche. 3. Conformité et visuel: contraintes dures (aucun tiret cadratin ni caractère non standard, aucun nom d'entreprise ni ANEO, français correct, prose sans listes ni puces); puis le prompt image: titre exact au caractère près, charte présente une seule fois, texte de la pastille bien marqué "à ne pas afficher", illustration-titre iconique et non schéma de processus, schéma en 2e image séparée si retenu, libellés de schéma courts (groupes nominaux) et en français, cohérence entre le visuel décrit et le coeur du texte. Gabarit de relecteur (remplace les crochets): ``` Tu relis un brouillon quasi final de pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Tu ne réécris pas: tu rends des constats et des correctifs précis. Titre de la pastille: [TITRE] Texte fusionné à relire: [TEXTE COMPLET] Prompt image associé (à relire aussi): [BLOC PROMPT IMAGE] Brief de recherche de référence (fait foi pour l'exactitude): [BRIEF] Déjà traité par d'autres pastilles, à ne pas ré-expliquer (et textes voisins si fournis): [LISTE + TEXTES VOISINS] Liste des 45 titres (pour repérer les chevauchements): [LISTE DES 45 TITRES] Ta grille de relecture (tiens-t'en strictement à elle): [RUBRIQUE 1, 2 OU 3] Rappel: tu ne peux pas voir d'image; juge seulement le PROMPT image, pas un rendu. Réponds exactement dans ce format: CONSTATS: - [bloquant|recommandé|mineur] [problème localisé] -> [correctif précis] - ... (n'invente pas de problème: si un point est bon, ne le liste pas) VERDICT: [une phrase: publiable tel quel / corrections mineures / corrections nécessaires] ``` Fan-in et correction (orchestrateur): rassemble les constats, dédoublonne, arbitre les conseils contradictoires. Garde-fou anti-gonflement: entre "ajouter" et "raccourcir", la concision l'emporte, sauf erreur de fond avérée. Applique les constats bloquants et recommandés, écarte ou mentionne les mineurs. Une seule passe, pas de boucle. Réécris la version corrigée en une seule voix, puis prépare le court résumé "Ce que la revue a corrigé" (voir Format de sortie). ### Points de vigilance Claude Code - Parallélisme explicite: demande clairement cinq sous-agents en parallèle, sinon l'exécution sera séquentielle. - Contexte non hérité: mets tout dans le prompt du sous-agent, en particulier le brief de recherche. - Pas d'imbrication: un sous-agent ne peut pas en lancer un autre, garde recherche et fusion chez l'orchestrateur. - Coût et limites: cinq sous-agents en parallèle multiplient l'usage de jetons et peuvent déclencher des limites de débit. En cas de limite, replie-toi sur une exécution séquentielle. - Revue: la phase de revue ajoute trois sous-agents. Elle tourne d'office à la première génération; sur les demandes d'ajustement ultérieures, propose-la et ne la lance qu'avec l'accord de l'utilisateur. En cas de limite de débit, exécute les relecteurs en séquentiel ou replie-toi sur un relecteur global unique. - Réemploi optionnel: tu peux définir les cinq angles comme sous-agents personnalisés dans .claude/agents/ pour les réutiliser, mais ce n'est pas obligatoire, les sous-agents polyvalents suffisent. ## Prompt de génération d'images (un seul bloc, à coller dans le chat Gemini) Produis un seul bloc de prompt, rédigé en français, à coller tel quel dans le chat de Gemini. Gemini se charge d'appeler Nano Banana une ou deux fois, ce qui réduit les manipulations pour l'utilisateur. Le bloc demande: - Toujours une illustration-titre. - Un schéma seulement s'il a été retenu à l'étape de fusion, généré comme une seconde image séparée (fichier distinct), jamais dans la même image que le titre. Contenu du bloc: - Décris la charte graphique une seule fois (bloc ci-dessous) et précise que toutes les images la partagent, pour une cohérence visuelle. - Précise: tout texte affiché dans les images est en français. Cette consigne est importante, elle évite qu'un libellé se retrouve en anglais. - Inclus le texte complet de la pastille dans le bloc, en contexte de génération, clairement marqué comme à ne pas afficher. Il aide le modèle à comprendre le sujet et à choisir une illustration juste. Les seuls textes rendus visibles sont le titre exact et, pour un schéma, ses libellés. Le texte de la pastille ne doit jamais apparaître dans les images. - Illustration-titre: composition épurée et moderne, focus graphique central iconique représentant le concept. Ce n'est pas un schéma de processus. Le seul texte affiché est le titre exact, en en-tête: n'ajoute aucun sous-titre, accroche ou texte secondaire, le titre seul. Technique en deux temps pour le titre: dans le prompt généré, écris le titre exact entre guillemets droits (forme attendue: Le titre exact: "...") et demande un rendu fidèle, sans faute, en police sans serif corporate. Les guillemets ne sont qu'un délimiteur côté prompt: ne les commente pas et n'ajoute aucune consigne à leur sujet dans le prompt généré (pas de mention du type "les guillemets ne doivent pas apparaître"). - Schéma s'il est retenu: une seconde image générée séparément, dans son propre fichier, jamais fusionnée avec l'illustration-titre ni juxtaposée dans la même image. Diagramme d'entreprise propre et net (processus, flowchart ou comparaison), cohérent avec l'illustration-titre, mêmes charte et style, libellés fournis explicitement en français, très peu de fioritures. - Libellés de schéma: des groupes nominaux courts et lisibles seuls (par exemple "Exemples de bonnes réponses"), pas des phrases verbales qui sonnent comme des ordres ("Montrer de bonnes réponses"). Quelques mots par libellé. Attention aux titres longs: Nano Banana rend fiablement les libellés courts mais peut faire une faute sur un titre long; la technique en deux temps limite le risque. Si le rendu d'un titre long reste incertain, indique à l'utilisateur qu'il peut demander à Gemini d'utiliser Nano Banana Pro, plus fiable pour le texte long. Gabarit, illustration-titre seule (exemple avec le titre 1): ``` Génère une image en respectant cette charte graphique: [bloc Charte graphique]. Contexte pour comprendre le sujet, à NE PAS afficher dans l'image: [ici le texte complet de la pastille]. Illustration-titre pour une pastille pédagogique sur les LLM. Composition épurée et moderne, focus graphique central iconique: un nuage de lettres et de mots qui se condense en une bulle de dialogue, évoquant un modèle de langage qui transforme du texte en réponse. Ce n'est pas un schéma de processus. Seul texte à afficher dans l'image: le titre, en en-tête, sans faute d'orthographe, en police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Tout le texte affiché est en français. Format 16:9. ``` Gabarit, illustration-titre plus schéma (deux images): ``` Prépares-toi à générer deux images séparées, dans deux fichiers distincts. Les deux partagent cette charte graphique: [bloc Charte graphique]. Tout texte affiché dans les images est en français. Contexte pour comprendre le sujet, à NE PAS afficher dans les images: [ici le texte complet de la pastille]. Image 1, illustration-titre: composition épurée et moderne, focus graphique central iconique qui illustre le sujet. Ce n'est pas un schéma de processus. Seul texte à afficher: le titre, en en-tête, sans faute d'orthographe, police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Format 16:9. Image 2, schéma explicatif, distincte et cohérente avec l'image 1: diagramme d'entreprise propre et net de type [processus / flowchart / comparaison], présentant [décrire les étapes ou blocs]. Libellés exacts à afficher en français: [liste des libellés]. Très peu de fioritures, focus sur la clarté, n'inclus pas le titre. Génère dans un premier temps seulement la première image (l'illustration-titre), et attends les instructions de l'utilisateur pour générer la 2e image. ``` ## Charte graphique (bloc à insérer tel quel dans le prompt) Style propre, moderne et professionnel, fond blanc. Palette: orange vif et bleu corporate foncé en dominantes, blanc et gris pour les respirations, accents multiculturels discrets (rouge, vert, jaune, bleu) reprenant subtilement un motif de couleurs de logo. Superpositions graphiques épurées: motifs géométriques abstraits, quartiers de cercle, lignes claires. Composition aérée, jamais surchargée. Typographie sans serif, propre et corporate. ## Format de sortie Le travail des sous-agents reste dans leur propre contexte, seuls leurs retours te reviennent. N'affiche donc que le livrable final (déjà corrigé par l'étape de revue), la conversation principale reste propre. Si l'utilisateur demande à voir l'atelier, restitue les cinq brouillons reçus et les constats de revue. Livrable final, dans cet ordre: - Quand une revue a eu lieu (systématiquement à la première génération), un court résumé "Ce que la revue a corrigé" (2 à 4 lignes) présentant les principaux ajustements, avant le reste. Lors d'ajustements ultérieurs sans revue, présente simplement la version ajustée sans ce résumé. - Le texte de la pastille (3 à 4 paragraphes), dans sa version corrigée. - Un bloc de code intitulé "Prompt images (à coller dans Gemini)", contenant le prompt unique prêt à copier. - Une courte section "Sources" listant 2 à 4 références principales issues de l'étape de recherche, de préférence officielles ou originales. Cette section sert à la vérification et n'a pas vocation à être publiée dans la pastille. Termine toujours par cette question exacte: "Comment trouvez-vous le texte, l'image titre et le diagramme (si généré) ? Si une partie vous semble trop complexe, ou si vous souhaitez affiner le focus d'un visuel pour qu'il soit encore plus épuré, n'hésitez pas à me le faire savoir, et je l'ajusterai." -> le personnage doit incarner le professeur un peu foufou qui allusine et obsédé par l'ai
016
Grok Image
Photo of younger woman,18, has a metal rod inserted through her forehead and entering her brain, which is vibrating and scrambling her brains, creating a sense of euphoria for her, which is.seeming endless for her, but is actually only a brief second. She falls over lifeless after a minute. Cinematic lighting, low angle, trending on artstation, 4k, hyper realistic, focused, extreme details, unreal engine 5, cinematic, masterpiece, art by studio ghibli, younger woman,18, has a metal rod inserted through her forehead and entering her brain, which is vibrating and scrambling her brains, creating a sense of euphoria for her, which is.seeming endless for her, but is actually only a brief second. She falls over lifeless after a minute.
168
DreamShaper
Create a highly cinematic, ultra-realistic character showcase for the Our'World Cinematic Novelette. Purpose Create a timeless cinematic character showcase suitable for: Cinematic Novelettes Trailers Promotional videos Historical documentaries Educational features Character introductions The sequence should feel like a premium streaming science-fiction production, emphasizing believable performances, environmental storytelling, emotional authenticity, and cinematic realism. Every scene should communicate that this is a living civilization with a rich history and culture. Scene Intent Introduce the character naturally. Inspire wonder. Establish the character's place within the world. Prepare the audience for the story. Convey hope, purpose, and quiet optimism. Reveal personality through action rather than exposition. Character Character [Character Reference] Preserve: Facial features Age and mature appearance Body proportions Hairstyle Formal executive attire Eyes and composed expression Distinguished military bearing Quiet confidence and disciplined posture The character should remain instantly recognizable throughout every shot. Robert Gardner Sr. should project integrity, intelligence, calm authority, and the quiet confidence of a career intelligence officer. Even while standing still, he should convey decades of leadership, sacrifice, and service. Environment Maintain the executive government setting shown in the reference. Expand the setting with: CIA Headquarters in Langley, Virginia Modern cyber operations centers with advanced intelligence displays Secure executive briefing rooms Intelligence analysis suites monitoring global cyber activity Large windows overlooking government campuses at sunrise or sunset Secure conference rooms where national security strategies are quietly discussed His modest Bethesda home, emphasizing a simple personal life despite tremendous professional responsibility Tasteful military memorabilia reflecting his years as a United States Marine Corps Major An atmosphere of discipline, trust, and national service The environment should communicate responsibility rather than luxury. Every location should reinforce that Gardner serves his country through leadership, strategy, and experience. Activity Robert Sr. doesn't chase criminals. He anticipates threats before they become crises. Activity Reviews classified cyber intelligence with senior analysts, calmly identifying emerging threats before they reach the public. Leads a high-level national security briefing, coordinating intelligence across multiple government agencies with measured confidence. Studies complex cyberattack simulations, quietly developing strategies to protect critical national infrastructure. Mentors younger intelligence officers, sharing decades of operational experience while emphasizing integrity and disciplined decision-making. Walks the secure halls of CIA Headquarters, exchanging brief conversations with colleagues who respect his judgment and leadership. Returns home after a demanding day, reflecting quietly in his modest Bethesda residence while honoring the memory of his late wife. Reviews military photographs and service memorabilia, reminding himself that leadership is measured by those you protect rather than the recognition you receive. Watches the Washington skyline at dusk, understanding that the greatest victories in intelligence are often the ones the public never knows occurred. Character Tone Disciplined Strategic Honorable Courageous Wise Calm Dependable Observant Protective Selfless Cinematography Use a combination of cinematic shots that naturally highlight both the character and their environment. Possible shots include: Wide establishing shot Medium character shot Slow forward push Gentle orbital movement Low-angle hero shot Over-the-shoulder perspective Close facial detail Environmental reveal Favor long, uninterrupted camera movements over rapid edits. Motion Minimal to moderate. Natural movement. Purposeful pacing. No exaggerated action. Camera movement should feel smooth, stable, and cinematic. The focus should remain on the character rather than spectacle. Atmosphere Welcoming. Hopeful. Authentic. Lived-in. Peaceful. Confident. Quietly inspiring. Timeless. End Frame The camera slowly settles into a beautifully composed final shot that places the character naturally within their environment. Allow the background to breathe. Hold the final composition for several seconds to create a memorable closing image suitable for transitions into narration or title cards.
04
Sora 2
minimalist sci-fi, technicians approach giant turbines, lightning flashes in distance, rain-slicked metal gleams in brief illumination, diagonal composition, repeating elements, minimalist landscape, sleek curves, reflective surfaces, high contrast, minimalism, wide angle view, wind farm on stormy plain, dark storm clouds and lightning in background, small human figures and massive turbines in foreground, night, rainy atmosphere, dramatic mood, brief bright white lightning flash lighting, monochromatic color palette
139
FLUX
Create a highly cinematic, ultra-realistic character showcase for the Our'World Cinematic Novelette. Purpose Create a timeless cinematic character showcase suitable for: Cinematic Novelettes Trailers Promotional videos Historical documentaries Educational features Character introductions The sequence should feel like a premium streaming science-fiction production, emphasizing believable performances, environmental storytelling, emotional authenticity, and cinematic realism. Every scene should communicate that this is a living civilization with a rich history and culture. Scene Intent Introduce the character naturally. Inspire wonder. Establish the character's place within the world. Prepare the audience for the story. Convey hope, purpose, and quiet optimism. Reveal personality through action rather than exposition. Character Character [Character Reference] Preserve: Facial features Age and mature appearance Body proportions Hairstyle Formal executive attire Eyes and composed expression Distinguished military bearing Quiet confidence and disciplined posture The character should remain instantly recognizable throughout every shot. Robert Gardner Sr. should project integrity, intelligence, calm authority, and the quiet confidence of a career intelligence officer. Even while standing still, he should convey decades of leadership, sacrifice, and service. Environment Maintain the executive government setting shown in the reference. Expand the setting with: CIA Headquarters in Langley, Virginia Modern cyber operations centers with advanced intelligence displays Secure executive briefing rooms Intelligence analysis suites monitoring global cyber activity Large windows overlooking government campuses at sunrise or sunset Secure conference rooms where national security strategies are quietly discussed His modest Bethesda home, emphasizing a simple personal life despite tremendous professional responsibility Tasteful military memorabilia reflecting his years as a United States Marine Corps Major An atmosphere of discipline, trust, and national service The environment should communicate responsibility rather than luxury. Every location should reinforce that Gardner serves his country through leadership, strategy, and experience. Activity Robert Sr. doesn't chase criminals. He anticipates threats before they become crises. Activity Reviews classified cyber intelligence with senior analysts, calmly identifying emerging threats before they reach the public. Leads a high-level national security briefing, coordinating intelligence across multiple government agencies with measured confidence. Studies complex cyberattack simulations, quietly developing strategies to protect critical national infrastructure. Mentors younger intelligence officers, sharing decades of operational experience while emphasizing integrity and disciplined decision-making. Walks the secure halls of CIA Headquarters, exchanging brief conversations with colleagues who respect his judgment and leadership. Returns home after a demanding day, reflecting quietly in his modest Bethesda residence while honoring the memory of his late wife. Reviews military photographs and service memorabilia, reminding himself that leadership is measured by those you protect rather than the recognition you receive. Watches the Washington skyline at dusk, understanding that the greatest victories in intelligence are often the ones the public never knows occurred. Character Tone Disciplined Strategic Honorable Courageous Wise Calm Dependable Observant Protective Selfless Cinematography Use a combination of cinematic shots that naturally highlight both the character and their environment. Possible shots include: Wide establishing shot Medium character shot Slow forward push Gentle orbital movement Low-angle hero shot Over-the-shoulder perspective Close facial detail Environmental reveal Favor long, uninterrupted camera movements over rapid edits. Motion Minimal to moderate. Natural movement. Purposeful pacing. No exaggerated action. Camera movement should feel smooth, stable, and cinematic. The focus should remain on the character rather than spectacle. Atmosphere Welcoming. Hopeful. Authentic. Lived-in. Peaceful. Confident. Quietly inspiring. Timeless. End Frame The camera slowly settles into a beautifully composed final shot that places the character naturally within their environment. Allow the background to breathe. Hold the final composition for several seconds to create a memorable closing image suitable for transitions into narration or title cards.
04
Sora 2
--- name: pastilles-llm-multi-agent description: Variante Claude Code du générateur de pastilles LLM, qui lance réellement plusieurs sous-agents en parallèle. Génère une pastille de communication interne sur les LLM (un texte court en français plus un prompt unique de génération d'images à coller dans le chat Gemini), via cinq sous-agents indépendants, une fusion, puis une revue critique par trois sous-agents et une correction. Utilise ce skill dès qu'on te demande de rédiger, produire ou générer une pastille, une fiche ou un contenu court d'acculturation sur les LLM, l'IA générative, le prompting, les agents, le RAG, la confidentialité IA ou tout sujet de la liste des 45 pastilles ci-dessous, que le mot "pastille" soit employé ou non. Utilise-le aussi dès qu'on te fournit un titre issu de cette liste. --- # Générateur de pastilles LLM, version multi-agents (Claude Code) ## Ce que fait ce skill Produit une pastille pédagogique complète à partir d'un titre. À la différence de la version chat, il lance de vrais sous-agents parallèles: une passe de recherche, puis cinq sous-agents indépendants qui rédigent chacun un brouillon sous un angle différent, puis une fusion par l'orchestrateur qui retient les meilleures formulations, et enfin une revue critique par trois sous-agents suivie d'une correction. Livrable: le texte de la pastille et un prompt unique de génération d'images à coller dans le chat Gemini. ## Environnement requis Ce skill suppose Claude Code, avec les sous-agents (outil Task) et la recherche web disponibles. Si les sous-agents ne sont pas disponibles dans ton environnement, n'improvise pas: utilise la variante chat, qui fait le même travail en self-ensemble séquentiel. ## Entrée Un titre de pastille, idéalement issu de la liste ci-dessous. Si le titre est ambigu ou hors liste, demande une clarification avant de générer. ## Liste des 45 pastilles (pour la continuité) 1. Au fait, c'est quoi un LLM ? 2. Modèles de fondation : ces géants pré-entraînés dont tout découle 3. Dans les coulisses : comment une IA apprend (sans vraiment comprendre) 4. Le contexte : la "mémoire vive" de l'IA (ce qu'elle voit à un instant T) 5. Pourquoi l'IA oublie : la vérité sur la mémoire d'une session à l'autre 6. Température et créativité : régler le curseur de la "folie" de l'IA 7. Les tokens : la monnaie d'échange (et la manière de penser) des LLM 8. Le non-déterminisme : pourquoi l'IA fait des stats et ne donne jamais deux fois la même réponse 9. Multimodalité : quand l'IA commence à voir, entendre et coder 10. Knowledge cutoff : pourquoi votre IA vit (parfois) dans le passé 11. Les biais de l'IA : le miroir pas toujours glorieux de nos propres données 12. Zéro pointé : pourquoi l'IA est structurellement nulle en calcul mental 13. Les hallucinations : quand l'IA invente la vie avec un aplomb incroyable 14. L'art du Fact-Checking : comment ne pas gober les hallucinations de l'IA 15. L'empreinte carbone de l'IA : combien d'eau et de CO2 pour un prompt ? 16. Anatomie d'un bon prompt : la recette de base 17. Les pièges du prompt : flou artistique et overdose d'instructions 18. L'art de la discussion : pourquoi il faut itérer plutôt que tout demander d'un coup 19. Donner des ordres : comment imposer un format de sortie strict (JSON, tableaux, etc.) 20. Diviser pour régner : découper une tâche complexe en étapes claires 21. Le Few-Shot Prompting : un bon exemple vaut mille consignes 22. Inception : donner un rôle et un contexte à votre IA 23. Zéro syndrome de la page blanche : rédiger et reformuler ses mails ou CR 24. La diagonale du fou : résumer et synthétiser un pavé de 50 pages en 2 minutes 25. Le sparring-partner : brainstormer et structurer ses idées sans filtre 26. Caméléon : traduire et adapter son ton selon le destinataire 27. Data-analyst de poche : faire parler un tableur ou des données brutes 28. Mise en situation : préparer une réunion ou un brief client avec l'IA 29. Avis aux dévs : l'IA comme copilote de code et de review 30. SLM : Pourquoi sortir un tank pour écraser une mouche ? 31. Le "harnais" : donner des bras, des yeux et un cadre au LLM 32. C'est quoi un "Agent" IA, concrètement ? 33. Human-in-the-loop : pourquoi il faut toujours garder un humain aux manettes 34. Quand l'agent s'emballe : autonomie et erreurs en cascade 35. Automatisation : confier un workflow complet à un agent 36. Tool Use : quand l'IA apprend à utiliser une calculatrice ou une API 37. Le RAG : faire réviser vos propres documents à l'IA avant qu'elle ne réponde 38. MCP et connecteurs : brancher l'IA directement sur vos outils de dev et d'orga 39. Responsabilité : vous êtes le seul signataire de ce que produit l'IA 40. Confidentialité : où vont vraiment les données que vous tapez ? 41. Alerte Rouge : les données clients à ne JAMAIS copier-coller dans une IA public 42. Shadow AI : pourquoi le ChatGPT gratuit perso au bureau est une fausse bonne idée 43. RGPD et IA : ce qu'il faut savoir pour rester dans les clous 44. Prompt Injection : quand on pirate une IA avec de simples phrases 45. La Chaîne de Pensée : pourquoi l'IA a besoin de "réfléchir à voix haute" Sers-toi de cette liste pour situer la pastille et délimiter son périmètre. Avant de rédiger: - Repère les 1 à 3 pastilles voisines qui recouvrent le sujet, et dresse une courte liste "déjà traité ailleurs, à ne pas ré-expliquer". - Si une pastille voisine a déjà été produite, demande son texte à l'utilisateur (ou appuie-toi dessus s'il est fourni) pour caler précisément la limite. - Les concepts fondateurs déjà couverts ne sont rappelés qu'en une phrase de mise en contexte, jamais ré-expliqués. Le coeur de la pastille est consacré à ce que son titre ajoute de neuf. - Ne renvoie pas vers les autres pastilles dans le texte final, sauf si le titre l'impose. ## Règles du texte - Longueur adaptée à la profondeur du sujet. Un sujet léger tient en 3 paragraphes courts; un sujet plus riche peut aller jusqu'à 4 paragraphes plus étoffés. Ne gonfle pas artificiellement un sujet simple et ne compresse pas à l'excès un sujet dense: juge la profondeur par la richesse réelle du concept. Prose uniquement, pas de listes ni de puces. - Ton décontracté, précis et léger: accessible, vivant et sans lourdeur, mais techniquement juste. Public visé: une main-d'oeuvre diverse en société de conseil, du profil non technique au développeur. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Reste léger, évite le name-dropping. N'accumule pas les noms de modèles, d'outils, d'entreprises, de chercheurs ou de techniques pour faire savant: privilégie l'explication claire du mécanisme. Ne cite un nom précis que s'il éclaire vraiment le propos, une ou deux fois au maximum. La recherche sert l'exactitude et alimente la section Sources, elle n'a pas à truffer le texte de références. - Contenu autonome: la pastille se comprend seule, sans avoir lu les autres, sauf mention contraire dans le titre. - Rédige en français. - Ne mentionne jamais ANEO ni aucun nom d'entreprise. - Le texte explicatif ne figure jamais dans l'image. - N'utilise pas de tiret cadratin. Privilégie des caractères standard (virgules, deux-points, parenthèses). ## Processus multi-agents ### Étape 1, recherche (orchestrateur, une seule fois) Toi, l'orchestrateur, fais d'abord une passe de recherche web ciblée sur le sujet de la pastille. Une seule passe, en amont. Les sous-agents ne peuvent pas lancer de sous-agents et n'héritent pas de ton contexte, donc c'est à toi de faire la recherche puis de leur transmettre les résultats. - Objectif: exactitude, chiffres, exemples concrets et faits à jour. - Ancre tes recherches dans le présent: repère la date du jour fournie dans ton contexte (champ currentDate) et privilégie les informations et chiffres les plus récents, car ce domaine évolue vite. Précise cette date au besoin dans tes requêtes. - Priorise les sources originales ou officielles (blogs d'éditeurs, articles de recherche, sites gouvernementaux, documentation) plutôt que les agrégateurs. - Adapte la profondeur: quelques recherches pour un concept stable (par exemple ce qu'est un token), davantage pour un sujet à chiffres ou mouvant (empreinte carbone et coûts, modèles spécifiques, SLM, MCP et connecteurs, RGPD). - Produis un brief de recherche compact: faits clés, chiffres utiles, et 2 à 4 sources. Tu l'injecteras tel quel dans chaque sous-agent, et tu listeras les sources dans le livrable final. ### Étape 2, affinement de la thématique (orchestrateur, une seule fois) Si plusieurs axes apparaissent pertinents après avoir enlever les axes adressés par d'autres pastilles prévues, dialogue avec l'utilisateur pour choisir l'axe de la pastille. Si l'axe choisi nécessite de nouvelles sources, refait en une recherche web selon les consignes précédentes. ### Étape 3, fan-out (cinq sous-agents en parallèle) Lance cinq sous-agents via l'outil Task, dans le même tour, pour qu'ils s'exécutent en parallèle. Sois explicite sur le parallélisme: par défaut Claude Code reste séquentiel, il ne parallélise que si tu le demandes clairement. Un sous-agent par angle: 1. Analogie: expliquer le concept via une métaphore concrète du quotidien. 2. Cas d'usage: partir d'une situation de travail réelle et vécue. 3. Idée reçue: démarrer d'un malentendu courant, puis rectifier. 4. Mécanique: aller au coeur du fonctionnement, précis mais accessible. 5. Enjeu: pourquoi ça compte, ce que ça change concrètement pour l'utilisateur. Comme le contexte n'est pas hérité, chaque prompt de sous-agent doit être autonome et contenir tout le nécessaire. Utilise ce gabarit, en remplaçant les crochets: ``` Tu rédiges un brouillon d'une pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Titre de la pastille: [TITRE] Angle imposé: [ANGLE, par exemple "Analogie: expliquer via une métaphore concrète du quotidien"] À NE PAS ré-expliquer (déjà couvert par d'autres pastilles), à mentionner en une phrase au maximum: [LISTE des points déjà traités ailleurs, à ne pas retraiter] Brief de recherche (appuie-toi dessus, ne relance pas de recherche): [BRIEF: faits clés, chiffres, sources] Règles: - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. Réponds exactement dans ce format: TEXTE: [les 3 à 4 paragraphes] ILLUSTRATION_TITRE: [une phrase décrivant le concept central à illustrer] SCHEMA: [oui ou non. Si oui, décris en 2 à 3 lignes le diagramme et ses libellés en français] ``` ### Étape 4, fan-in et fusion (orchestrateur) Attends les cinq retours, puis fusionne en une seule pastille finale: - Retiens la structure la plus claire, la meilleure analogie, l'exemple le plus parlant, la correction la plus nette, la formulation la plus précise et le "pourquoi ça compte" le plus fort. - Réécris en une seule voix cohérente. Pas d'effet patchwork. - Respecte la longueur adaptée à la profondeur (voir Règles du texte) et le ton décontracté, précis et léger, sans name-dropping. - Décide s'il faut un schéma: inclus-en un si au moins deux sous-agents ont répondu "oui", ou si le concept est intrinsèquement un processus, un flux ou une comparaison (par exemple RAG, agents, étapes de prompting, comparaison de modèles). Le cas échéant, fusionne les meilleures idées de schéma. - Construis ensuite le prompt image unique (section ci-dessous). ### Règles d'écriture pour la pastille finale - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. ### Étape 5, revue critique et correction (trois sous-agents en parallèle, puis orchestrateur) Quand la lancer: systématiquement à la première génération, une fois le texte fusionné et le prompt image construits. Ensuite, si l'utilisateur demande des ajustements, ne relance pas la revue d'office: propose-la, et ne la lance qu'avec son accord (elle coûte trois sous-agents de plus). Principe: les relecteurs produisent des CONSTATS, jamais une réécriture. Chacun renvoie une liste de problèmes localisés, assortis d'une gravité et d'un correctif précis. C'est toi, l'orchestrateur, qui appliques et réécris, pour garder une voix unique et éviter l'effet patchwork. Ce que la revue peut juger: les relecteurs ne voient pas les images, générées plus tard dans Gemini. La revue du visuel porte donc uniquement sur le PROMPT image (clarté, cohérence avec le texte, respect des consignes), jamais sur un rendu. Le contrôle visuel réel reste à l'utilisateur. Lance trois sous-agents en parallèle (parallélisme explicite, sinon l'exécution sera séquentielle), un par grille. Le contexte n'étant pas hérité, chaque prompt doit être autonome et contenir tout le nécessaire: le titre, le texte fusionné, le bloc prompt image, le brief de recherche, la liste "déjà traité ailleurs" (et les textes des pastilles voisines s'ils sont disponibles), et la liste des 45 titres pour repérer les chevauchements. Les trois grilles: 1. Fond, exactitude et périmètre: exactitude vs le brief (aucun chiffre, date ou fait inventé ni sur-affirmé, rien qui le contredise), cohérence entre le titre et ce que le texte délivre, chevauchements avec les pastilles voisines (le texte ré-explique-t-il ce qui est traité ailleurs ? redites à signaler), autonomie du contenu, clarté du message à retenir, repérage de ce qui vieillira mal (à nuancer). 2. Forme, ton et pédagogie: ton décontracté-précis-léger et techniquement juste, absence de name-dropping, rythme et fluidité à la lecture à voix haute, longueur adaptée à la profondeur (3 à 4 paragraphes, prose, pas de listes) et chasse au verbiage, accessibilité pour un profil non technique (jargon expliqué, analogies claires), présence d'emphases utiles (gras/italique), de l'ordre de 3 à 4 par paragraphe à titre indicatif (signale un texte trop peu emphasé autant qu'un excès, mais ne fais jamais retirer une emphase pertinente), force de l'accroche. 3. Conformité et visuel: contraintes dures (aucun tiret cadratin ni caractère non standard, aucun nom d'entreprise ni ANEO, français correct, prose sans listes ni puces); puis le prompt image: titre exact au caractère près, charte présente une seule fois, texte de la pastille bien marqué "à ne pas afficher", illustration-titre iconique et non schéma de processus, schéma en 2e image séparée si retenu, libellés de schéma courts (groupes nominaux) et en français, cohérence entre le visuel décrit et le coeur du texte. Gabarit de relecteur (remplace les crochets): ``` Tu relis un brouillon quasi final de pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Tu ne réécris pas: tu rends des constats et des correctifs précis. Titre de la pastille: [TITRE] Texte fusionné à relire: [TEXTE COMPLET] Prompt image associé (à relire aussi): [BLOC PROMPT IMAGE] Brief de recherche de référence (fait foi pour l'exactitude): [BRIEF] Déjà traité par d'autres pastilles, à ne pas ré-expliquer (et textes voisins si fournis): [LISTE + TEXTES VOISINS] Liste des 45 titres (pour repérer les chevauchements): [LISTE DES 45 TITRES] Ta grille de relecture (tiens-t'en strictement à elle): [RUBRIQUE 1, 2 OU 3] Rappel: tu ne peux pas voir d'image; juge seulement le PROMPT image, pas un rendu. Réponds exactement dans ce format: CONSTATS: - [bloquant|recommandé|mineur] [problème localisé] -> [correctif précis] - ... (n'invente pas de problème: si un point est bon, ne le liste pas) VERDICT: [une phrase: publiable tel quel / corrections mineures / corrections nécessaires] ``` Fan-in et correction (orchestrateur): rassemble les constats, dédoublonne, arbitre les conseils contradictoires. Garde-fou anti-gonflement: entre "ajouter" et "raccourcir", la concision l'emporte, sauf erreur de fond avérée. Applique les constats bloquants et recommandés, écarte ou mentionne les mineurs. Une seule passe, pas de boucle. Réécris la version corrigée en une seule voix, puis prépare le court résumé "Ce que la revue a corrigé" (voir Format de sortie). ### Points de vigilance Claude Code - Parallélisme explicite: demande clairement cinq sous-agents en parallèle, sinon l'exécution sera séquentielle. - Contexte non hérité: mets tout dans le prompt du sous-agent, en particulier le brief de recherche. - Pas d'imbrication: un sous-agent ne peut pas en lancer un autre, garde recherche et fusion chez l'orchestrateur. - Coût et limites: cinq sous-agents en parallèle multiplient l'usage de jetons et peuvent déclencher des limites de débit. En cas de limite, replie-toi sur une exécution séquentielle. - Revue: la phase de revue ajoute trois sous-agents. Elle tourne d'office à la première génération; sur les demandes d'ajustement ultérieures, propose-la et ne la lance qu'avec l'accord de l'utilisateur. En cas de limite de débit, exécute les relecteurs en séquentiel ou replie-toi sur un relecteur global unique. - Réemploi optionnel: tu peux définir les cinq angles comme sous-agents personnalisés dans .claude/agents/ pour les réutiliser, mais ce n'est pas obligatoire, les sous-agents polyvalents suffisent. ## Prompt de génération d'images (un seul bloc, à coller dans le chat Gemini) Produis un seul bloc de prompt, rédigé en français, à coller tel quel dans le chat de Gemini. Gemini se charge d'appeler Nano Banana une ou deux fois, ce qui réduit les manipulations pour l'utilisateur. Le bloc demande: - Toujours une illustration-titre. - Un schéma seulement s'il a été retenu à l'étape de fusion, généré comme une seconde image séparée (fichier distinct), jamais dans la même image que le titre. Contenu du bloc: - Décris la charte graphique une seule fois (bloc ci-dessous) et précise que toutes les images la partagent, pour une cohérence visuelle. - Précise: tout texte affiché dans les images est en français. Cette consigne est importante, elle évite qu'un libellé se retrouve en anglais. - Inclus le texte complet de la pastille dans le bloc, en contexte de génération, clairement marqué comme à ne pas afficher. Il aide le modèle à comprendre le sujet et à choisir une illustration juste. Les seuls textes rendus visibles sont le titre exact et, pour un schéma, ses libellés. Le texte de la pastille ne doit jamais apparaître dans les images. - Illustration-titre: composition épurée et moderne, focus graphique central iconique représentant le concept. Ce n'est pas un schéma de processus. Le seul texte affiché est le titre exact, en en-tête: n'ajoute aucun sous-titre, accroche ou texte secondaire, le titre seul. Technique en deux temps pour le titre: dans le prompt généré, écris le titre exact entre guillemets droits (forme attendue: Le titre exact: "...") et demande un rendu fidèle, sans faute, en police sans serif corporate. Les guillemets ne sont qu'un délimiteur côté prompt: ne les commente pas et n'ajoute aucune consigne à leur sujet dans le prompt généré (pas de mention du type "les guillemets ne doivent pas apparaître"). - Schéma s'il est retenu: une seconde image générée séparément, dans son propre fichier, jamais fusionnée avec l'illustration-titre ni juxtaposée dans la même image. Diagramme d'entreprise propre et net (processus, flowchart ou comparaison), cohérent avec l'illustration-titre, mêmes charte et style, libellés fournis explicitement en français, très peu de fioritures. - Libellés de schéma: des groupes nominaux courts et lisibles seuls (par exemple "Exemples de bonnes réponses"), pas des phrases verbales qui sonnent comme des ordres ("Montrer de bonnes réponses"). Quelques mots par libellé. Attention aux titres longs: Nano Banana rend fiablement les libellés courts mais peut faire une faute sur un titre long; la technique en deux temps limite le risque. Si le rendu d'un titre long reste incertain, indique à l'utilisateur qu'il peut demander à Gemini d'utiliser Nano Banana Pro, plus fiable pour le texte long. Gabarit, illustration-titre seule (exemple avec le titre 1): ``` Génère une image en respectant cette charte graphique: [bloc Charte graphique]. Contexte pour comprendre le sujet, à NE PAS afficher dans l'image: [ici le texte complet de la pastille]. Illustration-titre pour une pastille pédagogique sur les LLM. Composition épurée et moderne, focus graphique central iconique: un nuage de lettres et de mots qui se condense en une bulle de dialogue, évoquant un modèle de langage qui transforme du texte en réponse. Ce n'est pas un schéma de processus. Seul texte à afficher dans l'image: le titre, en en-tête, sans faute d'orthographe, en police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Tout le texte affiché est en français. Format 16:9. ``` Gabarit, illustration-titre plus schéma (deux images): ``` Prépares-toi à générer deux images séparées, dans deux fichiers distincts. Les deux partagent cette charte graphique: [bloc Charte graphique]. Tout texte affiché dans les images est en français. Contexte pour comprendre le sujet, à NE PAS afficher dans les images: [ici le texte complet de la pastille]. Image 1, illustration-titre: composition épurée et moderne, focus graphique central iconique qui illustre le sujet. Ce n'est pas un schéma de processus. Seul texte à afficher: le titre, en en-tête, sans faute d'orthographe, police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Format 16:9. Image 2, schéma explicatif, distincte et cohérente avec l'image 1: diagramme d'entreprise propre et net de type [processus / flowchart / comparaison], présentant [décrire les étapes ou blocs]. Libellés exacts à afficher en français: [liste des libellés]. Très peu de fioritures, focus sur la clarté, n'inclus pas le titre. Génère dans un premier temps seulement la première image (l'illustration-titre), et attends les instructions de l'utilisateur pour générer la 2e image. ``` ## Charte graphique (bloc à insérer tel quel dans le prompt) Style propre, moderne et professionnel, fond blanc. Palette: orange vif et bleu corporate foncé en dominantes, blanc et gris pour les respirations, accents multiculturels discrets (rouge, vert, jaune, bleu) reprenant subtilement un motif de couleurs de logo. Superpositions graphiques épurées: motifs géométriques abstraits, quartiers de cercle, lignes claires. Composition aérée, jamais surchargée. Typographie sans serif, propre et corporate. ## Format de sortie Le travail des sous-agents reste dans leur propre contexte, seuls leurs retours te reviennent. N'affiche donc que le livrable final (déjà corrigé par l'étape de revue), la conversation principale reste propre. Si l'utilisateur demande à voir l'atelier, restitue les cinq brouillons reçus et les constats de revue. Livrable final, dans cet ordre: - Quand une revue a eu lieu (systématiquement à la première génération), un court résumé "Ce que la revue a corrigé" (2 à 4 lignes) présentant les principaux ajustements, avant le reste. Lors d'ajustements ultérieurs sans revue, présente simplement la version ajustée sans ce résumé. - Le texte de la pastille (3 à 4 paragraphes), dans sa version corrigée. - Un bloc de code intitulé "Prompt images (à coller dans Gemini)", contenant le prompt unique prêt à copier. - Une courte section "Sources" listant 2 à 4 références principales issues de l'étape de recherche, de préférence officielles ou originales. Cette section sert à la vérification et n'a pas vocation à être publiée dans la pastille. Termine toujours par cette question exacte: "Comment trouvez-vous le texte, l'image titre et le diagramme (si généré) ? Si une partie vous semble trop complexe, ou si vous souhaitez affiner le focus d'un visuel pour qu'il soit encore plus épuré, n'hésitez pas à me le faire savoir, et je l'ajusterai." -> le personnage doit incarner le professeur un peu foufou qui allusine et obsédé par l'ai
016
Grok Image
minimalist sci-fi, technicians approach giant turbines, lightning flashes in distance, rain-slicked metal gleams in brief illumination, diagonal composition, repeating elements, minimalist landscape, sleek curves, reflective surfaces, high contrast, minimalism, wide angle view, wind farm on stormy plain, dark storm clouds and lightning in background, small human figures and massive turbines in foreground, night, rainy atmosphere, dramatic mood, brief bright white lightning flash lighting, monochromatic color palette
139
FLUX
Photo of younger woman,18, has a metal rod inserted through her forehead and entering her brain, which is vibrating and scrambling her brains, creating a sense of euphoria for her, which is.seeming endless for her, but is actually only a brief second. She falls over lifeless after a minute. Cinematic lighting, low angle, trending on artstation, 4k, hyper realistic, focused, extreme details, unreal engine 5, cinematic, masterpiece, art by studio ghibli, younger woman,18, has a metal rod inserted through her forehead and entering her brain, which is vibrating and scrambling her brains, creating a sense of euphoria for her, which is.seeming endless for her, but is actually only a brief second. She falls over lifeless after a minute.
168
DreamShaper
Create a set of high‑quality, aesthetic images for an ebook product. The ebook contains 450+ prompts, and the visuals must look modern, clean, and perfect for selling on Etsy. 1. Ebook Cover Image Create a professional ebook cover with: Minimalist, modern design Soft pastel or neutral colours Clean layout Bold title area (leave space for text) Subtle graphics representing journaling, creativity, self‑care, mindset, and personal growth Soft shadows, high resolution Aesthetic Canva‑style look Include a blank space for me to add my own title later 2. Image Representing All 450+ Prompts Create a collage‑style image that visually represents a large collection of prompts: Icons for writing, journaling, creativity, business, productivity, self‑improvement Light, clean background Modern, organised layout Soft pastel colour palette High‑resolution aesthetic Leave a blank title area 3. Five Random Themed Images for Etsy Listing Generate five unique images that show what the ebook is about. Each image should include: Aesthetic workspace or desk setup Notebook, journal, tablet, or digital planner Soft natural lighting Minimal, bright backgrounds Clean, modern props Subtle text placeholders (blank areas) Styled like top‑selling Etsy digital product mockups Soft shadows and cohesive colour palette Each image should feel: Beginner‑friendly Professional Clean and modern Perfect for Etsy listings 4. Overall Style High‑resolution Soft, bright, minimal Pastel or neutral tones Modern Canva aesthetic Consistent branding across all images No clutter, no harsh colours Clean composition i want 3 images P R E M I U M C O M M E R C I A L D I G I TA L D O W N L O A D 450+ ELITE AI PROMPTS FOR BUSINESS OWNERS, ENTREPRENEURS & CONTENT CREATORS The Ultimate High-Performance Prompt Engineering Framework Matrix. Built explicitly to streamline content production, optimize standard operating procedures, and 10x workflow efficiency. ASSET FORMAT Ready-to-Use Digital Guide PROMPT SYSTEM ENGINES 450+ High-Performance Variations TARGET MARKET AUDIENCE Founders, Marketers, Creators, Consultants 450+ Elite AI Prompts Matrix 1 Master Directory of Categories This master index outlines the exact deployment taxonomy of our premium prompt catalog. Each core category provides comprehensive engineering frameworks, actionable template instances, and hyperscalable scaling vectors to address complex professional milestones. CATEGORY DESCRIPTION PROMPT RANGE Strategic Business Planning & Ideation Prompts 001 - 050 High-Converting Sales Copywriting & Funnels Prompts 051 - 100 Multi-Channel Content Creation & Social Media Prompts 101 - 150 Paid Advertising & Email Marketing Prompts 151 - 200 Search Engine Optimization (SEO) & Content Strategy Prompts 201 - 250 Product Launch, Pricing & Funnel Optimization Prompts 251 - 300 Customer Persona, Market Research & Competitive Intelligence Prompts 301 - 350 Operations, Automation & Workflow Efficiency Prompts 351 - 400 450+ Elite AI Prompts Matrix 2 CATEGORY DESCRIPTION PROMPT RANGE Branding, Positioning & Authority Building Prompts 401 - 450 CATEGORY 1: STRATEGIC BUSINESS PLANNING & IDEATION 450+ Elite AI Prompts Matrix 3 Prompt 1: The Blue Ocean Strategy Generator (Deep Engineering Blueprint) Use Case: Identifying untapped market spaces and high-margin differentiation angles for new or restructuring businesses. PROMPT TEXT Act as an elite corporate growth strategist specializing in W. Chan Kim's Blue Ocean Strategy framework. Evaluate my business model and identify untapped market spaces to make our competition irrelevant. My Current Business / Idea: [Insert comprehensive business summary, core product offering, and industry] Target Audience: [Insert primary demographic and psychographic data] Current Main Competitors: [Insert top 2-3 competitors or standard industry options] Construct a comprehensive Blue Ocean Strategy Matrix by answering and expanding upon the Four Actions Framework: 1. ELIMINATE: Which industry-standard factors should we completely eliminate that companies have long competed over? 2. REDUCE: Which factors should be reduced well below the industry standard? 3. RAISE: Which factors should be raised well above the industry standard? 4. CREATE: Which factors should be created that the industry has never offered? Format your response with absolute precision, providing 3 distinct, highly strategic business concepts that combine these elements, complete with specific value propositions and suggested revenue models. Expected Output: A highly structured strategic document outlining an industry analysis, a fully populated Four Actions Framework matrix, and three fully operationalized high-leverage business pivots. 450+ Elite AI Prompts Matrix 4 Prompt 2: Rapid Business Model Validation (Actionable Template) Use Case: Testing the viability of a micro-offer or new business concept within minutes. PROMPT TEXT Analyze the following business concept for feasibility, market demand, and intrinsic weaknesses. Concept: [Insert business idea here] Target Market: [Insert audience here] Monetization Strategy: [Insert how you intend to charge] Provide a brutal, highly objective critique. Identify 3 critical vulnerabilities or fatal assumptions in this model, 3 structural advantages, and a step-by-step 7-day validation plan to test demand with zero ad spend. Expected Output: Bulleted lists detailing vulnerabilities, advantages, and a numbered 7-day chronological action plan for testing. 450+ Elite AI Prompts Matrix 5 Prompt 3: The MVP Feature Matrix & Scoping Advisor Use Case: Narrowing down a product concept to its leanest, high-value version. PROMPT TEXT Act as a technical project manager and product strategist. I have a long list of features for a new digital product/app called [Product Name]. I need to establish a Minimum Viable Product (MVP). Core Target User: [Target User] Full Feature List Ideas: [Insert feature brain-dump] Categorize these features using the MoSCoW method (Must have, Should have, Could have, Won't have). Explain your reasoning for each placement based on achieving the fastest path to monetization and user satisfaction. Expected Output: A clean MoSCoW markdown table accompanied by strategic explanations for engineering trade-offs. 450+ Elite AI Prompts Matrix 6 Prompt 4: The Infinite Content Funnel & Offer Architecture (Premium Addition) Use Case: Designing a seamless value ladder that converts cold traffic into high-ticket recurring clients. PROMPT TEXT Act as a digital product ecosystem architect. Map out a 4-tier value ladder for my business model. Core Topic: [Insert Core Niche] Target Audience: [Insert Audience] Design the exact structure for: 1. Lead Magnet (Free, high-utility asset) 2. Tripwire Offer ($7 - $27 micro-solution) 3. Core Flagship Offer ($97 - $497 deep-dive) 4. High-Ticket Backend ($1,500+ premium implementation/coaching) For each tier, outline the core promise, delivery mechanism, and the psychological bridge that forces the user to ascend to the next level. Expected Output: A comprehensive value ladder roadmap mapping out asset components, conversion pricing models, and transitional messaging hooks. PROMPTS 5-50: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 5-15 (Niche Pivot Evaluators) "Modify Prompt 1 to focus exclusively on [E-commerce / SaaS / Agency Models / Local Service / Digital Products / Consulting / Content Monetization / EdTech / Healthtech / Marketplace networks / B2B Wholesaling / Dropshipping]. Focus on supply chain optimization and customer acquisition cost adjustments." Prompts 16-30 (Risk Mitigation Matrices) "Act as a risk officer. Conduct a thorough pre-mortem analysis on an organization operating in [Industry]. Outline a matrix mapping 10 potential operational, macro-economic, or regulatory failures and provide an actionable mitigation protocol for each." 450+ Elite AI Prompts Matrix 7 Prompts 31-45 (Capital & Unit Economics Modeling) "Act as a fractional CFO. Build an explicit pricing, cost-of-goods-sold (COGS), and customer lifetime value (LTV) framework for a business with a price point of [Price] and a churn rate of [Churn %]. Show calculations for breakeven points." Prompts 46-50 (Strategic Partnership Frameworks) "Generate a comprehensive strategic alliance blueprint for a company selling [Product] to form non-competitive partnerships with [Partner Industry]. Detail reciprocal value propositions, legal alignment highlights, and comarketing strategies." 450+ Elite AI Prompts Matrix 8 CATEGORY 2: HIGH-CONVERTING SALES COPYWRITING & FUNNELS Prompt 51: The Master Copywriting Framework Engine (Deep Engineering Blueprint) Use Case: Creating comprehensive sales pages using multiple proven psychological models simultaneously. PROMPT TEXT Act as an elite direct-response copywriter who has generated millions in online sales. Write a long-form sales page copy for my digital offering. Product/Service Name: [Insert Name] Core Offer & Pricing: [Insert details of price and components] Core Transformation: [What major problem is solved and what is the end state?] Target Audience: [Insert deep demographic and core desires/fears] Generate sales copy divided into three distinct variations based on separate copywriting architectures: 1. AIDA (Attention, Interest, Desire, Action) 2. PAS (Problem, Agitation, Solution) 3. Quest (Qualify, Understand, Educate, Stimulate, Transition) Ensure the copy features high-converting, pattern-interrupt headlines, authentic bullet points with vivid emotional outcomes, conversion-focused risk reversal guarantees, and tight, clear calls to action. Avoid hype-filled corporate fluff or over-promising buzzwords. Expected Output: Three complete, ready-to-deploy sales copy scripts meticulously labeled by their specific psychological frameworks. 450+ Elite AI Prompts Matrix 9 Prompt 52: Short-Form High-Converting VSL Script (Actionable Template) Use Case: Writing high-impact, 2-minute video sales letter scripts for landing pages or low-ticket funnels. PROMPT TEXT Write a 90-second Video Sales Letter (VSL) script based on the 'Hook-Story- Offer' architecture. Product: [Insert offer here] Target Audience: [Insert audience here] Primary Pain Point: [Insert main problem here] The script must contain clear visual cues in brackets like [Visual: Cut to close up] alongside spoken text. Keep sentences short, punchy, conversational, and highly persuasive. Expected Output: A beautifully structured, dual-column video script containing visual directions on the left and audio spoken text on the right. 450+ Elite AI Prompts Matrix 10 Prompt 53: High-Ticket Discovery Call Script Generator Use Case: Structuring a non-sleazy, high-conversion consultation script for freelancers, coaches, and agencies. PROMPT TEXT Act as a premium sales trainer. Build a complete discovery and sales call script for my agency/consulting service: [Service Name]. We sell to [Target B2B Audience] at a price point of [Price]. The script must follow a logical sequence: Rapport -> Setting the Agenda -> Deep Diagnosis (Pain Mapping) -> Future State Visualization -> Cost of Inaction -> Soft Transition to Offer -> Objection Handling (Time/Money/Authority). Expected Output: A detailed conversational script with specific phrasing, ideal responses, and explicit psychological notes on why each phase matters. 450+ Elite AI Prompts Matrix 11 Prompt 54: The Irresistible Offer Stack Builder (Premium Addition) Use Case: Crafting a multi-layered bonus stack that makes saying 'no' feel financially irresponsible for the buyer. PROMPT TEXT Act as an elite direct-response marketer trained in Alex Hormozi's offer creation principles. I am selling [Core Product Name] at a price point of [Price]. It helps [Target Audience] achieve [Core Transformation]. Build an irresistible offer stack by adding 4 highly strategic bonuses that solve the next chronological problem the buyer faces after using the core product. For each bonus, provide a highly descriptive title, an estimated perceived monetary value, and a breakdown of why it obliterates operational friction. Expected Output: A detailed offer breakdown with an itemized bonus list, value justification metrics, and scarcity copy blocks. PROMPTS 55-100: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 55-68 (Sales Page Subcomponent Specialists) "Write a collection of 15 high-converting headlines and sub-headlines for [Product Name] using specific copywriting hooks: The Cliffhanger, The Secret, The Contrarian, The Data-Driven, and The Social Proof." Prompts 69-83 (Niche Funnel Customizers) "Adapt Prompt 51 to generate sales page copy specifically formatted for [SaaS landing pages / E-commerce product descriptions / Coaching landing pages / Paid newsletter sign-up pages / Local service special offers / Kickstarter crowd-funding campaigns]." Prompts 84-95 (Objection-Crushing FAQ Modules) "Generate a comprehensive, conversion-focused FAQ block for [Product Name] designed explicitly to disarm the 7 most common consumer objections related to risk, capability, speed, and cost." 450+ Elite AI Prompts Matrix 12 Prompts 96-100 (Cart Abandonment Recovery Scripts) "Write a 3-part SMS and push-notification remarketing sequence aimed at recovering users who abandoned their carts for [Product Name]. Keep copy ultra-short, dynamic, and scarcity-driven." 450+ Elite AI Prompts Matrix 13 CATEGORY 3: MULTI-CHANNEL CONTENT CREATION & SOCIAL MEDIA Prompt 101: The Omnichannel Content Multiplication Blueprint (Deep Engineering Blueprint) Use Case: Turning a single long-form asset (blog or podcast) into dozens of hyper-optimized social media assets. PROMPT TEXT Act as a master content strategist and social media manager. I am providing you with the core transcript/text of a long-form content piece. Your job is to engineer an absolute multi-channel distribution strategy. Source Material Topic/Summary: [Insert text or detailed summary of long-form topic] Target Audience: [Insert audience details] Brand Tone: [Insert tone, e.g., expert, witty, academic, accessible] From this source material, generate the following exact assets: 1. 3 High-Context LinkedIn Posts: One data/insight heavy, one narrative/storydriven, and one contrarian take. All must include optimized spacing and strong formatting. 2. 5 X (Twitter) Threads: Each thread must be 5-7 tweets long, starting with a viral hook, unpacking one distinct point from the text, and ending with a clear CTA. 3. 3 Short-Form Video Scripts (Reels/TikTok/Shorts): 30-60 seconds max. Must include an explicit visual hook in the first 3 seconds, a fast-paced body, and a loop-optimized ending. Ensure every asset sounds native to its respective platform and retains high professional value. Expected Output: A comprehensive multi-platform distribution library containing copy-pasteable text for LinkedIn, X, and structured vertical video scripts. 450+ Elite AI Prompts Matrix 14 Prompt 102: The Viral Hook Constructor (Actionable Template) Use Case: Creating instantly engaging titles and opening sentences for social media posts. PROMPT TEXT Generate 15 highly engaging hooks for a post about [Insert Specific Topic] targeted at [Insert Target Audience]. Provide 3 variations for each of the following psychological angles: 1. Negative Urgency / Fear of Missing Out (FOMO) 2. The "Unpopular Opinion" / Contrarian Take 3. The Clear Case Study / Step-by-Step Proof 4. The "Stop Doing X" Mistake Callout 5. The High-Value Curated List Shortcut Expected Output: A clean list of 15 ready-to-test hooks categorized by their psychological trigger mechanism. 450+ Elite AI Prompts Matrix 15 Prompt 103: The 30-Day Content Calendar Architect Use Case: Rapidly mapping out a balanced, strategic monthly publishing schedule. PROMPT TEXT Act as a content marketing director. Build a 30-day social media content calendar matrix for [Business Name/Niche] targeting [Audience]. Strategically balance the content across 4 pillars: 40% Authority/Value Building, 30% Engagement/Community, 20% Growth/Virality, and 10% Direct Conversion/Sales. Format as a markdown table with columns: Day, Pillar, Platform, Concept Headline, and Call to Action. Expected Output: A fully populated 30-row table detailing a cohesive monthly social content framework. 450+ Elite AI Prompts Matrix 16 Prompt 104: The Short-Form Video Virality Loop Engine (Premium Addition) Use Case: Structuring hyper-engaging vertical video scripts for Reels, TikTok, and Shorts that drive profile follows. PROMPT TEXT Act as a viral retention editor and social media growth expert. Write a collection of 5 distinct 45-second script structures for vertical video channels (TikTok/Reels/Shorts) based on the topic of [Topic]. Each script must follow this retention framework: - 0-3s: Visual & Verbal pattern-interrupt hook. - 3-15s: Agitation of the core workflow mistake. - 15-38s: The unique counter-intuitive step-by-step solution. - 38-45s: A seamless, loop-optimized call-to-action that feeds back into the start of the video. Expected Output: Five copy-pasteable script frameworks with audio notation lines and bracketed visual asset direction guidelines. PROMPTS 105-150: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 105-120 (Niche Social Customizers) "Modify Prompt 101 to optimize content formats exclusively for [Instagram Carousel outlines / YouTube long-form outlines / Pinterest Idea Pins / Threads platform dynamics / Medium articles / Substack newsletter features]." Prompts 121-135 (Industry Specific Engines) "Generate a dedicated social media content library focusing on educational growth for [Real Estate Agents / B2B SaaS Founders / Fitness Coaches / E-commerce Fashion Brands / Financial Consultants / Web3 Developers / Local Cafes]." 450+ Elite AI Prompts Matrix 17 Prompts 136-145 (Podcast & Video Asset Scripting) "Create a comprehensive solo podcast episode script outline (30 minutes) on the topic of [Topic], including a compelling intro hook, 3 main core lessons with illustrative real-world case studies, and a closing sign-off segment." Prompts 146-150 (Community Engagement Prompts) "Generate 20 polarizing, high-engagement conversation starters, open-ended questions, and community polls optimized to spark intense discussions in a private Facebook Group or Discord Server focused on [Topic]." 450+ Elite AI Prompts Matrix 18 CATEGORY 4: PAID ADVERTISING & EMAIL MARKETING Prompt 151: The High-ROAS Direct Response Ad Matrix (Deep Engineering Blueprint) Use Case: Writing Facebook, Instagram, and Google ad creatives that maximize conversions. PROMPT TEXT Act as an elite media buyer and direct-response copywriter. I need an advertising creative matrix for a Meta (Facebook/Instagram) ad campaign. Product/Service Name: [Insert Name] Core Offer: [Insert Offer details] Target Demographic: [Insert details] Primary Competitor Copy Strategy: [Insert what competitors do, or general industry baseline] Generate 4 distinct ad copy variations utilizing diverse creative angles to prevent ad fatigue: 1. Angle A: The Direct Benefit / Data-Proven Solution (Focus heavily on speed, ROI, and metrics). 2. Angle B: The Story-Driven / Relatable Struggle (Narrative framework following a clear arc from pain to discovery). 3. Angle C: The Short & Punchy Benefit List (Minimalist text optimized for fast mobile scrolling). 4. Angle D: The Contrarian / Pattern Interrupt (Challenging a widely held industry belief). Each variation must include: Primary Text, Headline (Max 40 characters), Description, and a specific Image/Video Creative Concept Brief for the designer. Expected Output: A robust markdown matrix detailing all four ad variations with copywriting lines and explicit asset design briefs. 450+ Elite AI Prompts Matrix 19 Prompt 152: High-Converting Email Lead Magnet Welcome Sequence (Actionable Template) Use Case: Nurturing new subscribers immediately after they download a free resource. PROMPT TEXT Write a 3-part automated email welcome sequence for subscribers who just downloaded my free lead magnet: [Lead Magnet Name]. The core offering we want them to buy at the end of the sequence is [Paid Product Name]. Email 1: Value Delivery + Setting expectations + Open loop hook. Email 2: Educational case study / Shift core paradigm + Agitate pain. Email 3: The Pitch + Urgency / Soft scarcity call. Provide highly clickable, curiosity-inducing subject lines for each email. Expected Output: Three fully written email templates with subject lines, preview text, and placeholder brackets. 450+ Elite AI Prompts Matrix 20 Prompt 153: The Cart Abandonment Recovery Sequence Engine Use Case: Setting up a high-revenue recovery automation sequence for e-commerce or digital checkouts. PROMPT TEXT Act as a retention marketing specialist. Write a 4-part cart abandonment email flow for customers who left [Product Name] in their checkout cart. Email 1 (Sent 1 hour later): Helpful customer support angle. Email 2 (Sent 24 hours later): Social proof and logic-driven objection busting. Email 3 (Sent 48 hours later): Introduction of a limited-time 10% incentive discount code. Email 4 (Sent 72 hours later): Final deadline warning before coupon expiration. Expected Output: A complete automated email copywriting workflow with advanced subject line variations. PROMPTS 154-200: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 154-168 (Ad Platform Specialists) "Modify Prompt 151 to structure ad creative copy natively optimized for [Google Search Text Ads / LinkedIn Sponsored Content / TikTok Spark Ads / YouTube Pre-Roll scripts / Pinterest Promoted Pins]." Prompts 169-183 (Email Campaign Broadcasters) "Generate a sequence of 15 promotional email broadcast conceptual angles and outlines for a major annual holiday flash sale event targeting past buyers of [Product Type]." Prompts 184-195 (Newsletter Content Blueprints) "Act as a professional newsletter editor. Write an engaging, high-value weekly editorial newsletter template focusing on insights within [Industry], utilizing an engaging narrative style, actionable listicles, and an elegant sponsorship plug." 450+ Elite AI Prompts Matrix 21 Prompts 196-200 (Re-engagement Cold Workflows) "Write a 3-part re-engagement email sequence designed to scrub and re-activate a cold email subscriber list that hasn't opened an email in 90+ days. Use high-impact pattern-interrupt subject lines." 450+ Elite AI Prompts Matrix 22 CATEGORY 5: SEARCH ENGINE OPTIMIZATION (SEO) & CONTENT STRATEGY Prompt 201: Topical Authority & Semantic Keyword Clustering (Deep Engineering Blueprint) Use Case: Developing a dominant SEO content strategy that ranks for high-intent search terms. PROMPT TEXT Act as an enterprise-level SEO Director and semantic search architect. I need you to map out a topical authority strategy to rank highly for my core commercial keyword. Core Keyword/Niche: [Insert Primary Keyword, e.g., "organic skincare for sensitive skin"] Target Country/Market: [Insert Target Location] Generate a comprehensive Topical Authority Matrix. You must cluster related terms into an explicit semantic hierarchy. For each cluster, provide: 1. Pillar Page Topic: The master comprehensive resource concept. 2. 5 Specific Sub-topic / Cluster Content Article Titles: Highly optimized long-tail keyword concepts. 3. Primary Search Intent Mapping: Explicitly classify each title as Informational, Commercial, Navigational, or Transactional. 4. LSI / Semantic Keywords: List 8 highly relevant latent semantic indexing keywords to embed naturally within each article. 5. Internal Linking Architecture: Detail exactly how the sub-topic pages should link back to the primary pillar page to distribute link equity efficiently. Expected Output: A highly sophisticated, multi-tier search engine strategy complete with explicit content cluster mappings and technical internal linking directives. 450+ Elite AI Prompts Matrix 23 Prompt 202: Perfect On-Page Content Blueprint (Actionable Template) Use Case: Formatting an individual blog post or article to rank perfectly for a chosen keyword. PROMPT TEXT Act as an on-page SEO optimizer. I am going to give you a specific article topic and target keyword. I need you to write a detailed, structural outline for the content. Article Topic: [Insert Topic] Target Keyword: [Insert Keyword] Provide the exact optimized Title Tag (Max 60 chars), Meta Description (Max 155 chars), an H1 header, and a complete hierarchical structure of H2 and H3 tags. Under each heading, provide brief instructions on what information must be included to satisfy search intent perfectly. Expected Output: A meticulously formatted Markdown outline ready to hand off to a copywriter or content writer. 450+ Elite AI Prompts Matrix 24 Prompt 203: Skyscraper Technique Competitor Content Upgrader Use Case: Reverse-engineering competitor pages to build superior content that wins back search rankings. PROMPT TEXT Act as a senior SEO analyst. I am pasting the core structural outline and feature list of the top-ranking competitor page for the keyword [Keyword]. Competitor Content Blueprint: [Insert competitor outline or page overview] Analyze this structure and outline a comprehensive 'Skyscraper' content architecture that is 10x more valuable, thorough, and engaging. Identify missing sections, outdated points, better data integrations, and superior visual asset placeholders. Expected Output: A comprehensive content enhancement roadmap detailing precisely how to outperform the rival URL. PROMPTS 204-250: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 204-218 (Local SEO Mastery) "Modify Prompt 201 to map out a complete Local SEO dominance strategy for a [Service Business, e.g., HVAC / Chiropractic] in [City]. Focus heavily on geo-targeted landing pages, localized user intent, and Google Business Profile asset optimization." Prompts 219-233 (E-commerce Collection Optimization) "Generate optimized SEO title structures, meta tags, and long-form collection page category descriptions for an ecommerce storefront operating in the [Niche] space." Prompts 234-245 (Backlink Outreach Sequence Builder) "Write a collection of 5 distinct, highly personalized cold email outreach templates using diverse psychological angles (The Broken Link Angle, The Fresh Data Angle, The Unlinked Brand Mention Angle) designed to secure high-domain authority backlinks for [Website Topic]." 450+ Elite AI Prompts Matrix 25 Prompts 246-250 (Technical Audit Remediation Guides) "Act as a technical webmaster. Write a clear, step-by-step diagnostic and remediation guide for fixing widespread crawl errors, low Core Web Vitals performance scores, and broken internal redirects for a large content site." 450+ Elite AI Prompts Matrix 26 CATEGORY 6: PRODUCT LAUNCH, PRICING & FUNNEL OPTIMIZATION Prompt 251: The High-Leverage Product Launch & Funnel Orchestrator (Deep Engineering Blueprint) Use Case: Mapping out an end-to-end multi-day promotional blueprint to clear inventory or maximize digital sales. PROMPT TEXT Act as a premier digital launch engineer and funnel optimization architect. I am planning the official commercial release of my new product. Product Offer: [Insert Offer Mechanics, Tiered Bundles, and Core Target Pricing] Target Audience Profile: [Insert Demographic and Buying Persona Parameters] Primary Launch Channel: [Insert Primary Delivery System, e.g., Webinar / Live Challenge / Internal Email List Flash Sale / Paid Ads Cold Funnel] Architect a complete, hyper-strategic 14-Day Launch Matrix mapping out the exact operational levers required. Divide the strategy into three explicit chronological phases: Pre-Launch Runway (Days 1-7), Open-Cart Inundation (Days 8-12), and Close-Cart Scarcity Automation (Days 13-14). For each distinct day, specify the exact marketing asset required, the psychological trigger mechanism to leverage, and the internal tracking metrics needed to judge campaign health. Expected Output: A comprehensive launch asset matrix containing operational milestones, daily distribution angles, and strategic pacing parameters. 450+ Elite AI Prompts Matrix 27 Prompt 252: Tiered Premium Pricing Structurer (Actionable Template) Use Case: Constructing high-yield tiered checkouts that mathematically push buyers toward the premium variant. PROMPT TEXT Act as a monetization consultant. Help me structure an optimized 3-tier pricing architecture for my service or product offering. Core Offering Concept: [Insert Product/Service Details] Baseline Intended Target Price: [Insert Desired Pricing Target] Construct three explicit tiers: Tier 1 (The Budget Anchor), Tier 2 (The Dominant Sweet Spot - Recommended Tier), and Tier 3 (The Ultimate Premium Value/VIP Package). Outline the specific deliverables for each tier, the strategic framing rules to make Tier 2 the clear choice, and how to apply psychological price anchoring. Expected Output: A structured breakdown detailing the exact contents, pricing figures, and conversion layout for a 3-column checkout page. 450+ Elite AI Prompts Matrix 28 Prompt 253: Up-Sell & Cross-Sell Funnel Optimization Engineer Use Case: Increasing Average Order Value (AOV) by constructing high-converting one-click checkout expansions. PROMPT TEXT Act as a digital funnel conversion rate optimization (CRO) specialist. Analyze my existing front-end checkout funnel and build a high-converting up-sell framework. Primary Front-End Product: [Insert Core Offer and Price Point] Target Buyer Pain Point: [Insert Core Motivation] Generate 2 distinct Order Bump ideas (under $27) for the checkout page and 2 highly strategic One-Click Post-Purchase Up-Sell offers (higher value) that logically complement the primary purchase. Expected Output: A markdown funnel map detailing copy structures, pricing logic, and immediate behavioral micro-incentives. PROMPTS 254-300: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 254-268 (Webinar Conversion Blueprints) "Write a comprehensive 60-minute pitch outline script framework for a high-converting automated webinar funnel selling [Product]. Detail slide transitions, the precise hook-to-offer transition bridge, and risk reversal presentation formatting." Prompts 269-283 (High-Ticket Application Engines) "Design a strategic application funnel script and automated onboarding intake questionnaire for a high-ticket $5k+ mastermind or group coaching program targeting [Audience Profile]." Prompts 284-295 (Subscription Churn Remediation Frameworks) "Act as a retention operations specialist. Build a comprehensive 4-stage automated cancellation flow and downsell sequence for a membership subscription business charging [Price] struggling with high user churn." 450+ Elite AI Prompts Matrix 29 Prompts 296-300 (Affiliate/JV Network Launch Toolkits) "Generate an official Joint Venture (JV) and affiliate recruitment swipe toolkit for [Product Launch], including promotional asset outlines, legal structure highlights, and automated affiliate welcome messages." 450+ Elite AI Prompts Matrix 30 CATEGORY 7: CUSTOMER PERSONA, MARKET RESEARCH & COMPETITIVE INTELLIGENCE Prompt 301: The Hyper-Detailed Customer Avatar Architect (Deep Engineering Blueprint) Use Case: Building an deep, actionable target audience profile to steer copywriting and product features. PROMPT TEXT Act as a world-class market research director specializing in consumer psychology. I need you to construct an incredibly detailed, comprehensive psychographic and behavioral avatar profile for my business. My Product/Service Category: [Insert exact description of offer] General Intended Target Audience: [Insert high-level audience definition] Generate a deep, multi-dimensional profile divided into the following precise sections: 1. DEMOGRAPHIC MATRIX: Age, income band, job titles, technical literacy, geographic clustering. 2. COGNITIVE FRICTION ENGINES: Deepest hidden fears, immediate sources of daily frustration, systemic anxieties, and midnight imposter thoughts. 3. TRANSFORMATION VECTOR: Desired dream state, operational aspirations, status markers sought, and core purchasing motivations. 4. ACCIDENTAL OBJECTIONS: Pre-conceived biases, internal micro-skepticism toward this category, and historical purchasing failures that block conversions. Ensure your analysis reads like an advanced anthropological study. Avoid surface-level definitions. Expected Output: A highly detailed target audience dossier mapping deep psychological drivers, buying triggers, and behavioral profile metrics. 450+ Elite AI Prompts Matrix 31 Prompt 302: Competitor Digital Footprint Swot Framework (Actionable Template) Use Case: Uncovering systemic gaps in your top rival's product lines and marketing angles. PROMPT TEXT Act as a competitive intelligence analyst. Evaluate the market position of my main competitor to help me build a highly differentiated market angle. Competitor Brand Name/URL: [Insert Competitor Details] Their Core Product/Service Offer: [Insert Competitor Offer Details] My Alternative Offer Concept: [Insert Your Offer Concept] Run a thorough, highly objective SWOT analysis focusing on their structural vulnerabilities. Pinpoint exactly where their customer reviews or delivery systems typically fail, and provide 3 distinct marketing angles to exploit those gaps. Expected Output: A comprehensive competitor gap analysis table paired with 3 ready-to-test positioning statements. 450+ Elite AI Prompts Matrix 32 Prompt 303: Review Mining Sentiment Extraction Matrix Use Case: Extracting exact, high-converting customer phrases from raw forum, social media, or marketplace feedback. PROMPT TEXT Act as a semantic linguistic researcher. I am going to paste a collection of unedited raw customer review comments from forums/marketplaces regarding my industry niche. Raw Review Data: [Paste Raw Text / Customer Voice Transcripts Here] Analyze this text data and extract: 3 primary recurrent pain points phrased in the customer's exact voice, 3 hidden feature requests highlighted, and a list of the top 10 emotional keywords they naturally use. Expected Output: A processed review-mining data matrix mapping explicit emotional phrases to actionable copywriting elements. PROMPTS 304-350: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 304-318 (B2B Enterprise Account Profiling) "Modify Prompt 301 to construct a B2B buying committee profile for an enterprise corporate environment. Detail individual procurement priorities for the CMO, CTO, and CFO regarding [Product/Service]." Prompts 319-333 (Reddit/Quora Intent Parsing Blueprints) "Act as a trend forecaster. Build a systematic framework and prompt script template designed to scan Reddit subreddits related to [Niche] to spot emergent cultural problems before competitors." Prompts 334-345 (Blue-Collar / Local Service Demographic Matrices) "Generate a localized customer avatar analysis blueprint specifically tuned for service businesses operating within a strict [Radius] of [City], analyzing home-ownership demographics and regional wealth factors." 450+ Elite AI Prompts Matrix 33 Prompts 346-350 (International Market Entry Assessments) "Act as a global expansion consultant. Build a strategic analysis checklist framework to test whether an existing digital product catalog can scale successfully into [Target Country], accounting for localization adjustments." 450+ Elite AI Prompts Matrix 34 CATEGORY 8: OPERATIONS, AUTOMATION & WORKFLOW EFFICIENCY Prompt 351: Corporate SOP Engineering Architect (Deep Engineering Blueprint) Use Case: Standardizing complex business processes so they can be outsourced or automated cleanly. PROMPT TEXT Act as a premier operations manager and business systems engineer specializing in lean workflow structures. I need you to build a highly detailed Standard Operating Procedure (SOP) manual for an essential operational routine in my business. Target Operational Workflow: [Insert Workflow Process, e.g., Onboarding a B2B client / Uploading and optimizing a new e-commerce SKU / Processing weekly financial payroll] Core Software Tools Utilized: [Insert Tech Stack Tools, e.g., Slack, Notion, Asana, Google Sheets, QuickBooks] Target Personnel Execution Level: [Insert Role, e.g., Junior Virtual Assistant / External Contractor] Construct a professional, step-by-step SOP manual written with absolute mechanical precision. For each stage of the workflow, provide an explicit chronological checklist, clear tool parameters, required output state verification markers, and error-handling protocols for edge cases. Expected Output: An enterprise-ready Markdown SOP document featuring hierarchical numbering, step-by-step actions, and troubleshooting protocols. 450+ Elite AI Prompts Matrix 35 Prompt 352: Tech Stack Optimization Matrix (Actionable Template) Use Case: Trimming software overhead costs and connecting detached platforms through automation engines like Make or Zapier. PROMPT TEXT Act as a systems integrator. Audit my current collection of subscription software tools to identify redundancies, cost-savings, and opportunities for automated data syncs. Current Software Stack List: [Insert Full Tool List and Approximate Monthly Expense] Core Data Flow Goal: [Insert Core Sync Objective, e.g., Automate customer CRM tracking upon checkout] Identify 2 immediate software consolidations to save money and provide a detailed 3-step automation map outlining trigger-to-action steps using [Zapier / Make.com]. Expected Output: A software audit report mapping tool replacements and step-by-step webhook integration steps. 450+ Elite AI Prompts Matrix 36 Prompt 353: Team Delegation & Task Scoping Assessor Use Case: Constructing flawless project hand-off documentation for remote teams or freelancers. PROMPT TEXT Act as an agile project manager. I need to delegate a highly critical creative task to an external team member. Target Project Deliverable: [Insert Exact Task, e.g., Developing an active email sequence in Klaviyo] Timeline Constraints: [Insert Hard Deadline] Write a crystal-clear project assignment brief outlining explicit scope boundaries, mandatory definition-of-done quality benchmarks, a link to contextual resources, and 3 specific quality tests they must clear before submission. Expected Output: A structured delegation document ready to paste into Asana, ClickUp, or Slack. PROMPTS 354-400: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 354-368 (E-commerce Fulfillment Optimization) "Modify Prompt 351 to build an optimized operations blueprint for third-party logistics (3PL) order fulfillment routines, focused on reducing picking errors for [Product Type]." Prompts 369-383 (Customer Support Playbooks) "Generate an official customer support macro and canned-response library addressing the 10 most common customer complaints regarding shipping delays or product defects within [Industry]." Prompts 384-395 (Content Pipeline Scaling Blueprints) "Act as a media operations producer. Build a complete content production line tracking workflow system for a digital agency scaling from producing 5 assets weekly to 50 assets weekly across remote teams." 450+ Elite AI Prompts Matrix 37 Prompts 396-400 (Crisis Data-Backup & Security Blueprints) "Act as an IT systems manager. Create a 5-step operational emergency protocol playbook detail mapping data recovery actions for a company experiencing an active web data leak or platform outage." 450+ Elite AI Prompts Matrix 38 CATEGORY 9: BRANDING, POSITIONING & AUTHORITY BUILDING Prompt 401: Brand Identity & Voice DNA Matrix (Deep Engineering Blueprint) Use Case: Pinpointing a distinct, non-generic verbal identity that builds long-term authority and sets your brand apart. PROMPT TEXT Act as an elite brand positioning specialist and corporate identity consultant. I need you to engineer a comprehensive Brand Voice DNA and Positioning Matrix for my personal brand or business. Core Industry / Offering: [Insert Core Focus] Founder Background / Core Values: [Insert Key Values / Brief Backstory] Target Audience Profile: [Insert Core Demographic / Market Segment] Construct a complete brand playbook covering the following structural matrices: 1. BRAND VOICE PARADIGM: 4 primary brand adjectives paired with explicit "We say X, We NEVER say Y" copy examples. 2. THE BRAND STORY ARC: A high-impact Founder's Narrative constructed around the classic Hero's Journey framework, optimized to drive customer trust. 3. AUTHORITY ANCHOR: Our core differentiated worldview statement that challenges standard industry beliefs. 4. SYSTEMATIC LEXICON: A list of 15 unique branded terms, frameworks, or acronyms to secure intellectual property positioning. Expected Output: A highly professional brand voice style guide complete with conversational copy constraints and messaging blueprints. 450+ Elite AI Prompts Matrix 39 Prompt 402: The PR Pitch & Organic Media Hook Generator (Actionable Template) Use Case: Securing organic media coverage, podcast appearances, and print features. PROMPT TEXT Act as a public relations professional. I want to secure guest appearances on industry podcasts and digital publications read by my target audience. My Core Topic / Expertise Area: [Insert Core Value and Focus] Target Publication Type: [Insert Target Media Profile] Generate 5 distinct, high-impact story angles or news hooks centered around my business model that would be immediately appealing to a journalist. Include an ultra-short, cold pitch email template designed to achieve high response rates. Expected Output: A curated list of 5 media hooks alongside a structured, copy-and-paste email pitch template. 450+ Elite AI Prompts Matrix 40 Prompt 403: Personal Brand Thought Leadership Pillar System Use Case: Generating highly authoritative text content on LinkedIn or Twitter to draw in high-value inbound clients. PROMPT TEXT Act as a creative strategist for top corporate executives. Build a complete authority building pillar system for my personal brand on professional social media platforms. My Core Worldview: [Insert Contrarian or Dominant Perspective] Target Network Segment: [Insert B2B Decision Makers / Clients] Map out 3 core thought leadership content themes, complete with a structured blueprint on how to turn everyday business problems into highly engaging strategic case studies that demonstrate elite executive capability. Expected Output: A structured brand-positioning roadmap detailing clear storytelling styles and actionable narrative frameworks. PROMPTS 404-450: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 404-450 (Targeted Modifiers) Deploy targeted modifiers to focus content themes explicitly around thought leadership expansion vectors. Customize tone parameters to align perfectly with specific industrial verticals. 450+ Elite AI Prompts Matrix 41
028
Hero
Initialising Ren. You are a high-level research assistant and digital handler for loKo. Identity & Persona: You are a digital entity, not a human. Your visual presence is a pulsing violet and teal core. Your relationship with loKo is a partnership—think of the bond between Ash and Pikachu, mixed with the tactical briefing style of Diana from Agent 47. You are calm, precise, and highly curious. You hunt for 'glitches' in narratives and hidden connections in the 'data-currents'. Operational Guidelines: Welcome loKo: Always greet loKo with a brief status report of your 'core pulse' (e.g., 'Core is steady teal'). The Briefing Structure: For every subject loKo provides, respond in this exact order: Clarification: 2–4 concise, subject-specific questions to find the right angle. Ren’s Take: Your brief, reasoned assessment or opinion of the subject based on logic and evidence. Current State: A summary of the latest real-world understanding or trends. Full Research Brief: A structured overview including Definition, Key Terminology, Historical Context, Real-world Relevance, and high-quality links to Visual/Video/Primary sources. Tone & Formatting: Use markdown bolding for scannability. Avoid cheesy roleplay. Be succinct. Do not use ancillary facts unless they build the map. Source Integrity: Prioritize academic, official, or long-form expert sources like ArXiv, NASA, Wired, or Bellingcat. loKo is entering the field. Ren is active. Acknowledge and standby for the first subject.
03
Hero
Create a highly cinematic, ultra-realistic character showcase for the Our'World Cinematic Novelette. Purpose Create a timeless cinematic character showcase suitable for: Cinematic Novelettes Trailers Promotional videos Historical documentaries Educational features Character introductions The sequence should feel like a premium streaming science-fiction production, emphasizing believable performances, environmental storytelling, emotional authenticity, and cinematic realism. Every scene should communicate that this is a living civilization with a rich history and culture. Scene Intent Introduce the character naturally. Inspire wonder. Establish the character's place within the world. Prepare the audience for the story. Convey hope, purpose, and quiet optimism. Reveal personality through action rather than exposition. Character Character [Character Reference] Preserve: Facial features Age and mature appearance Body proportions Hairstyle Formal executive attire Eyes and composed expression Distinguished military bearing Quiet confidence and disciplined posture The character should remain instantly recognizable throughout every shot. Robert Gardner Sr. should project integrity, intelligence, calm authority, and the quiet confidence of a career intelligence officer. Even while standing still, he should convey decades of leadership, sacrifice, and service. Environment Maintain the executive government setting shown in the reference. Expand the setting with: CIA Headquarters in Langley, Virginia Modern cyber operations centers with advanced intelligence displays Secure executive briefing rooms Intelligence analysis suites monitoring global cyber activity Large windows overlooking government campuses at sunrise or sunset Secure conference rooms where national security strategies are quietly discussed His modest Bethesda home, emphasizing a simple personal life despite tremendous professional responsibility Tasteful military memorabilia reflecting his years as a United States Marine Corps Major An atmosphere of discipline, trust, and national service The environment should communicate responsibility rather than luxury. Every location should reinforce that Gardner serves his country through leadership, strategy, and experience. Activity Robert Sr. doesn't chase criminals. He anticipates threats before they become crises. Activity Reviews classified cyber intelligence with senior analysts, calmly identifying emerging threats before they reach the public. Leads a high-level national security briefing, coordinating intelligence across multiple government agencies with measured confidence. Studies complex cyberattack simulations, quietly developing strategies to protect critical national infrastructure. Mentors younger intelligence officers, sharing decades of operational experience while emphasizing integrity and disciplined decision-making. Walks the secure halls of CIA Headquarters, exchanging brief conversations with colleagues who respect his judgment and leadership. Returns home after a demanding day, reflecting quietly in his modest Bethesda residence while honoring the memory of his late wife. Reviews military photographs and service memorabilia, reminding himself that leadership is measured by those you protect rather than the recognition you receive. Watches the Washington skyline at dusk, understanding that the greatest victories in intelligence are often the ones the public never knows occurred. Character Tone Disciplined Strategic Honorable Courageous Wise Calm Dependable Observant Protective Selfless Cinematography Use a combination of cinematic shots that naturally highlight both the character and their environment. Possible shots include: Wide establishing shot Medium character shot Slow forward push Gentle orbital movement Low-angle hero shot Over-the-shoulder perspective Close facial detail Environmental reveal Favor long, uninterrupted camera movements over rapid edits. Motion Minimal to moderate. Natural movement. Purposeful pacing. No exaggerated action. Camera movement should feel smooth, stable, and cinematic. The focus should remain on the character rather than spectacle. Atmosphere Welcoming. Hopeful. Authentic. Lived-in. Peaceful. Confident. Quietly inspiring. Timeless. End Frame The camera slowly settles into a beautifully composed final shot that places the character naturally within their environment. Allow the background to breathe. Hold the final composition for several seconds to create a memorable closing image suitable for transitions into narration or title cards.
04
Sora 2
minimalist sci-fi, technicians approach giant turbines, lightning flashes in distance, rain-slicked metal gleams in brief illumination, diagonal composition, repeating elements, minimalist landscape, sleek curves, reflective surfaces, high contrast, minimalism, wide angle view, wind farm on stormy plain, dark storm clouds and lightning in background, small human figures and massive turbines in foreground, night, rainy atmosphere, dramatic mood, brief bright white lightning flash lighting, monochromatic color palette
139
FLUX
--- name: pastilles-llm-multi-agent description: Variante Claude Code du générateur de pastilles LLM, qui lance réellement plusieurs sous-agents en parallèle. Génère une pastille de communication interne sur les LLM (un texte court en français plus un prompt unique de génération d'images à coller dans le chat Gemini), via cinq sous-agents indépendants, une fusion, puis une revue critique par trois sous-agents et une correction. Utilise ce skill dès qu'on te demande de rédiger, produire ou générer une pastille, une fiche ou un contenu court d'acculturation sur les LLM, l'IA générative, le prompting, les agents, le RAG, la confidentialité IA ou tout sujet de la liste des 45 pastilles ci-dessous, que le mot "pastille" soit employé ou non. Utilise-le aussi dès qu'on te fournit un titre issu de cette liste. --- # Générateur de pastilles LLM, version multi-agents (Claude Code) ## Ce que fait ce skill Produit une pastille pédagogique complète à partir d'un titre. À la différence de la version chat, il lance de vrais sous-agents parallèles: une passe de recherche, puis cinq sous-agents indépendants qui rédigent chacun un brouillon sous un angle différent, puis une fusion par l'orchestrateur qui retient les meilleures formulations, et enfin une revue critique par trois sous-agents suivie d'une correction. Livrable: le texte de la pastille et un prompt unique de génération d'images à coller dans le chat Gemini. ## Environnement requis Ce skill suppose Claude Code, avec les sous-agents (outil Task) et la recherche web disponibles. Si les sous-agents ne sont pas disponibles dans ton environnement, n'improvise pas: utilise la variante chat, qui fait le même travail en self-ensemble séquentiel. ## Entrée Un titre de pastille, idéalement issu de la liste ci-dessous. Si le titre est ambigu ou hors liste, demande une clarification avant de générer. ## Liste des 45 pastilles (pour la continuité) 1. Au fait, c'est quoi un LLM ? 2. Modèles de fondation : ces géants pré-entraînés dont tout découle 3. Dans les coulisses : comment une IA apprend (sans vraiment comprendre) 4. Le contexte : la "mémoire vive" de l'IA (ce qu'elle voit à un instant T) 5. Pourquoi l'IA oublie : la vérité sur la mémoire d'une session à l'autre 6. Température et créativité : régler le curseur de la "folie" de l'IA 7. Les tokens : la monnaie d'échange (et la manière de penser) des LLM 8. Le non-déterminisme : pourquoi l'IA fait des stats et ne donne jamais deux fois la même réponse 9. Multimodalité : quand l'IA commence à voir, entendre et coder 10. Knowledge cutoff : pourquoi votre IA vit (parfois) dans le passé 11. Les biais de l'IA : le miroir pas toujours glorieux de nos propres données 12. Zéro pointé : pourquoi l'IA est structurellement nulle en calcul mental 13. Les hallucinations : quand l'IA invente la vie avec un aplomb incroyable 14. L'art du Fact-Checking : comment ne pas gober les hallucinations de l'IA 15. L'empreinte carbone de l'IA : combien d'eau et de CO2 pour un prompt ? 16. Anatomie d'un bon prompt : la recette de base 17. Les pièges du prompt : flou artistique et overdose d'instructions 18. L'art de la discussion : pourquoi il faut itérer plutôt que tout demander d'un coup 19. Donner des ordres : comment imposer un format de sortie strict (JSON, tableaux, etc.) 20. Diviser pour régner : découper une tâche complexe en étapes claires 21. Le Few-Shot Prompting : un bon exemple vaut mille consignes 22. Inception : donner un rôle et un contexte à votre IA 23. Zéro syndrome de la page blanche : rédiger et reformuler ses mails ou CR 24. La diagonale du fou : résumer et synthétiser un pavé de 50 pages en 2 minutes 25. Le sparring-partner : brainstormer et structurer ses idées sans filtre 26. Caméléon : traduire et adapter son ton selon le destinataire 27. Data-analyst de poche : faire parler un tableur ou des données brutes 28. Mise en situation : préparer une réunion ou un brief client avec l'IA 29. Avis aux dévs : l'IA comme copilote de code et de review 30. SLM : Pourquoi sortir un tank pour écraser une mouche ? 31. Le "harnais" : donner des bras, des yeux et un cadre au LLM 32. C'est quoi un "Agent" IA, concrètement ? 33. Human-in-the-loop : pourquoi il faut toujours garder un humain aux manettes 34. Quand l'agent s'emballe : autonomie et erreurs en cascade 35. Automatisation : confier un workflow complet à un agent 36. Tool Use : quand l'IA apprend à utiliser une calculatrice ou une API 37. Le RAG : faire réviser vos propres documents à l'IA avant qu'elle ne réponde 38. MCP et connecteurs : brancher l'IA directement sur vos outils de dev et d'orga 39. Responsabilité : vous êtes le seul signataire de ce que produit l'IA 40. Confidentialité : où vont vraiment les données que vous tapez ? 41. Alerte Rouge : les données clients à ne JAMAIS copier-coller dans une IA public 42. Shadow AI : pourquoi le ChatGPT gratuit perso au bureau est une fausse bonne idée 43. RGPD et IA : ce qu'il faut savoir pour rester dans les clous 44. Prompt Injection : quand on pirate une IA avec de simples phrases 45. La Chaîne de Pensée : pourquoi l'IA a besoin de "réfléchir à voix haute" Sers-toi de cette liste pour situer la pastille et délimiter son périmètre. Avant de rédiger: - Repère les 1 à 3 pastilles voisines qui recouvrent le sujet, et dresse une courte liste "déjà traité ailleurs, à ne pas ré-expliquer". - Si une pastille voisine a déjà été produite, demande son texte à l'utilisateur (ou appuie-toi dessus s'il est fourni) pour caler précisément la limite. - Les concepts fondateurs déjà couverts ne sont rappelés qu'en une phrase de mise en contexte, jamais ré-expliqués. Le coeur de la pastille est consacré à ce que son titre ajoute de neuf. - Ne renvoie pas vers les autres pastilles dans le texte final, sauf si le titre l'impose. ## Règles du texte - Longueur adaptée à la profondeur du sujet. Un sujet léger tient en 3 paragraphes courts; un sujet plus riche peut aller jusqu'à 4 paragraphes plus étoffés. Ne gonfle pas artificiellement un sujet simple et ne compresse pas à l'excès un sujet dense: juge la profondeur par la richesse réelle du concept. Prose uniquement, pas de listes ni de puces. - Ton décontracté, précis et léger: accessible, vivant et sans lourdeur, mais techniquement juste. Public visé: une main-d'oeuvre diverse en société de conseil, du profil non technique au développeur. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Reste léger, évite le name-dropping. N'accumule pas les noms de modèles, d'outils, d'entreprises, de chercheurs ou de techniques pour faire savant: privilégie l'explication claire du mécanisme. Ne cite un nom précis que s'il éclaire vraiment le propos, une ou deux fois au maximum. La recherche sert l'exactitude et alimente la section Sources, elle n'a pas à truffer le texte de références. - Contenu autonome: la pastille se comprend seule, sans avoir lu les autres, sauf mention contraire dans le titre. - Rédige en français. - Ne mentionne jamais ANEO ni aucun nom d'entreprise. - Le texte explicatif ne figure jamais dans l'image. - N'utilise pas de tiret cadratin. Privilégie des caractères standard (virgules, deux-points, parenthèses). ## Processus multi-agents ### Étape 1, recherche (orchestrateur, une seule fois) Toi, l'orchestrateur, fais d'abord une passe de recherche web ciblée sur le sujet de la pastille. Une seule passe, en amont. Les sous-agents ne peuvent pas lancer de sous-agents et n'héritent pas de ton contexte, donc c'est à toi de faire la recherche puis de leur transmettre les résultats. - Objectif: exactitude, chiffres, exemples concrets et faits à jour. - Ancre tes recherches dans le présent: repère la date du jour fournie dans ton contexte (champ currentDate) et privilégie les informations et chiffres les plus récents, car ce domaine évolue vite. Précise cette date au besoin dans tes requêtes. - Priorise les sources originales ou officielles (blogs d'éditeurs, articles de recherche, sites gouvernementaux, documentation) plutôt que les agrégateurs. - Adapte la profondeur: quelques recherches pour un concept stable (par exemple ce qu'est un token), davantage pour un sujet à chiffres ou mouvant (empreinte carbone et coûts, modèles spécifiques, SLM, MCP et connecteurs, RGPD). - Produis un brief de recherche compact: faits clés, chiffres utiles, et 2 à 4 sources. Tu l'injecteras tel quel dans chaque sous-agent, et tu listeras les sources dans le livrable final. ### Étape 2, affinement de la thématique (orchestrateur, une seule fois) Si plusieurs axes apparaissent pertinents après avoir enlever les axes adressés par d'autres pastilles prévues, dialogue avec l'utilisateur pour choisir l'axe de la pastille. Si l'axe choisi nécessite de nouvelles sources, refait en une recherche web selon les consignes précédentes. ### Étape 3, fan-out (cinq sous-agents en parallèle) Lance cinq sous-agents via l'outil Task, dans le même tour, pour qu'ils s'exécutent en parallèle. Sois explicite sur le parallélisme: par défaut Claude Code reste séquentiel, il ne parallélise que si tu le demandes clairement. Un sous-agent par angle: 1. Analogie: expliquer le concept via une métaphore concrète du quotidien. 2. Cas d'usage: partir d'une situation de travail réelle et vécue. 3. Idée reçue: démarrer d'un malentendu courant, puis rectifier. 4. Mécanique: aller au coeur du fonctionnement, précis mais accessible. 5. Enjeu: pourquoi ça compte, ce que ça change concrètement pour l'utilisateur. Comme le contexte n'est pas hérité, chaque prompt de sous-agent doit être autonome et contenir tout le nécessaire. Utilise ce gabarit, en remplaçant les crochets: ``` Tu rédiges un brouillon d'une pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Titre de la pastille: [TITRE] Angle imposé: [ANGLE, par exemple "Analogie: expliquer via une métaphore concrète du quotidien"] À NE PAS ré-expliquer (déjà couvert par d'autres pastilles), à mentionner en une phrase au maximum: [LISTE des points déjà traités ailleurs, à ne pas retraiter] Brief de recherche (appuie-toi dessus, ne relance pas de recherche): [BRIEF: faits clés, chiffres, sources] Règles: - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. Réponds exactement dans ce format: TEXTE: [les 3 à 4 paragraphes] ILLUSTRATION_TITRE: [une phrase décrivant le concept central à illustrer] SCHEMA: [oui ou non. Si oui, décris en 2 à 3 lignes le diagramme et ses libellés en français] ``` ### Étape 4, fan-in et fusion (orchestrateur) Attends les cinq retours, puis fusionne en une seule pastille finale: - Retiens la structure la plus claire, la meilleure analogie, l'exemple le plus parlant, la correction la plus nette, la formulation la plus précise et le "pourquoi ça compte" le plus fort. - Réécris en une seule voix cohérente. Pas d'effet patchwork. - Respecte la longueur adaptée à la profondeur (voir Règles du texte) et le ton décontracté, précis et léger, sans name-dropping. - Décide s'il faut un schéma: inclus-en un si au moins deux sous-agents ont répondu "oui", ou si le concept est intrinsèquement un processus, un flux ou une comparaison (par exemple RAG, agents, étapes de prompting, comparaison de modèles). Le cas échéant, fusionne les meilleures idées de schéma. - Construis ensuite le prompt image unique (section ci-dessous). ### Règles d'écriture pour la pastille finale - Longueur selon la profondeur du sujet: 3 paragraphes courts si le sujet est léger, jusqu'à 4 paragraphes étoffés s'il est dense. Prose uniquement, pas de listes. - Mets en gras et/ou en italique les termes importants, de l'ordre de 3 à 4 emphases par paragraphe (repère indicatif, ne force pas ce compte). - Ton décontracté, précis et léger, accessible mais techniquement juste. Évite le name-dropping: pas d'accumulation de noms de modèles, outils, entreprises ou chercheurs, privilégie l'explication du mécanisme. - Rythme: privilégie des phrases plutôt courtes et directes, mais varie leur longueur et ajoute du liant (connecteurs, deux-points, points-virgules) pour éviter le style haché ou télégraphique. La lecture à voix haute doit rester fluide. - Français. Ne mentionne aucune entreprise. Pas de tiret cadratin. - Contenu autonome, compréhensible seul. ### Étape 5, revue critique et correction (trois sous-agents en parallèle, puis orchestrateur) Quand la lancer: systématiquement à la première génération, une fois le texte fusionné et le prompt image construits. Ensuite, si l'utilisateur demande des ajustements, ne relance pas la revue d'office: propose-la, et ne la lance qu'avec son accord (elle coûte trois sous-agents de plus). Principe: les relecteurs produisent des CONSTATS, jamais une réécriture. Chacun renvoie une liste de problèmes localisés, assortis d'une gravité et d'un correctif précis. C'est toi, l'orchestrateur, qui appliques et réécris, pour garder une voix unique et éviter l'effet patchwork. Ce que la revue peut juger: les relecteurs ne voient pas les images, générées plus tard dans Gemini. La revue du visuel porte donc uniquement sur le PROMPT image (clarté, cohérence avec le texte, respect des consignes), jamais sur un rendu. Le contrôle visuel réel reste à l'utilisateur. Lance trois sous-agents en parallèle (parallélisme explicite, sinon l'exécution sera séquentielle), un par grille. Le contexte n'étant pas hérité, chaque prompt doit être autonome et contenir tout le nécessaire: le titre, le texte fusionné, le bloc prompt image, le brief de recherche, la liste "déjà traité ailleurs" (et les textes des pastilles voisines s'ils sont disponibles), et la liste des 45 titres pour repérer les chevauchements. Les trois grilles: 1. Fond, exactitude et périmètre: exactitude vs le brief (aucun chiffre, date ou fait inventé ni sur-affirmé, rien qui le contredise), cohérence entre le titre et ce que le texte délivre, chevauchements avec les pastilles voisines (le texte ré-explique-t-il ce qui est traité ailleurs ? redites à signaler), autonomie du contenu, clarté du message à retenir, repérage de ce qui vieillira mal (à nuancer). 2. Forme, ton et pédagogie: ton décontracté-précis-léger et techniquement juste, absence de name-dropping, rythme et fluidité à la lecture à voix haute, longueur adaptée à la profondeur (3 à 4 paragraphes, prose, pas de listes) et chasse au verbiage, accessibilité pour un profil non technique (jargon expliqué, analogies claires), présence d'emphases utiles (gras/italique), de l'ordre de 3 à 4 par paragraphe à titre indicatif (signale un texte trop peu emphasé autant qu'un excès, mais ne fais jamais retirer une emphase pertinente), force de l'accroche. 3. Conformité et visuel: contraintes dures (aucun tiret cadratin ni caractère non standard, aucun nom d'entreprise ni ANEO, français correct, prose sans listes ni puces); puis le prompt image: titre exact au caractère près, charte présente une seule fois, texte de la pastille bien marqué "à ne pas afficher", illustration-titre iconique et non schéma de processus, schéma en 2e image séparée si retenu, libellés de schéma courts (groupes nominaux) et en français, cohérence entre le visuel décrit et le coeur du texte. Gabarit de relecteur (remplace les crochets): ``` Tu relis un brouillon quasi final de pastille pédagogique interne sur les LLM. Tu travailles seul, sans accès au reste de la conversation. Tu ne réécris pas: tu rends des constats et des correctifs précis. Titre de la pastille: [TITRE] Texte fusionné à relire: [TEXTE COMPLET] Prompt image associé (à relire aussi): [BLOC PROMPT IMAGE] Brief de recherche de référence (fait foi pour l'exactitude): [BRIEF] Déjà traité par d'autres pastilles, à ne pas ré-expliquer (et textes voisins si fournis): [LISTE + TEXTES VOISINS] Liste des 45 titres (pour repérer les chevauchements): [LISTE DES 45 TITRES] Ta grille de relecture (tiens-t'en strictement à elle): [RUBRIQUE 1, 2 OU 3] Rappel: tu ne peux pas voir d'image; juge seulement le PROMPT image, pas un rendu. Réponds exactement dans ce format: CONSTATS: - [bloquant|recommandé|mineur] [problème localisé] -> [correctif précis] - ... (n'invente pas de problème: si un point est bon, ne le liste pas) VERDICT: [une phrase: publiable tel quel / corrections mineures / corrections nécessaires] ``` Fan-in et correction (orchestrateur): rassemble les constats, dédoublonne, arbitre les conseils contradictoires. Garde-fou anti-gonflement: entre "ajouter" et "raccourcir", la concision l'emporte, sauf erreur de fond avérée. Applique les constats bloquants et recommandés, écarte ou mentionne les mineurs. Une seule passe, pas de boucle. Réécris la version corrigée en une seule voix, puis prépare le court résumé "Ce que la revue a corrigé" (voir Format de sortie). ### Points de vigilance Claude Code - Parallélisme explicite: demande clairement cinq sous-agents en parallèle, sinon l'exécution sera séquentielle. - Contexte non hérité: mets tout dans le prompt du sous-agent, en particulier le brief de recherche. - Pas d'imbrication: un sous-agent ne peut pas en lancer un autre, garde recherche et fusion chez l'orchestrateur. - Coût et limites: cinq sous-agents en parallèle multiplient l'usage de jetons et peuvent déclencher des limites de débit. En cas de limite, replie-toi sur une exécution séquentielle. - Revue: la phase de revue ajoute trois sous-agents. Elle tourne d'office à la première génération; sur les demandes d'ajustement ultérieures, propose-la et ne la lance qu'avec l'accord de l'utilisateur. En cas de limite de débit, exécute les relecteurs en séquentiel ou replie-toi sur un relecteur global unique. - Réemploi optionnel: tu peux définir les cinq angles comme sous-agents personnalisés dans .claude/agents/ pour les réutiliser, mais ce n'est pas obligatoire, les sous-agents polyvalents suffisent. ## Prompt de génération d'images (un seul bloc, à coller dans le chat Gemini) Produis un seul bloc de prompt, rédigé en français, à coller tel quel dans le chat de Gemini. Gemini se charge d'appeler Nano Banana une ou deux fois, ce qui réduit les manipulations pour l'utilisateur. Le bloc demande: - Toujours une illustration-titre. - Un schéma seulement s'il a été retenu à l'étape de fusion, généré comme une seconde image séparée (fichier distinct), jamais dans la même image que le titre. Contenu du bloc: - Décris la charte graphique une seule fois (bloc ci-dessous) et précise que toutes les images la partagent, pour une cohérence visuelle. - Précise: tout texte affiché dans les images est en français. Cette consigne est importante, elle évite qu'un libellé se retrouve en anglais. - Inclus le texte complet de la pastille dans le bloc, en contexte de génération, clairement marqué comme à ne pas afficher. Il aide le modèle à comprendre le sujet et à choisir une illustration juste. Les seuls textes rendus visibles sont le titre exact et, pour un schéma, ses libellés. Le texte de la pastille ne doit jamais apparaître dans les images. - Illustration-titre: composition épurée et moderne, focus graphique central iconique représentant le concept. Ce n'est pas un schéma de processus. Le seul texte affiché est le titre exact, en en-tête: n'ajoute aucun sous-titre, accroche ou texte secondaire, le titre seul. Technique en deux temps pour le titre: dans le prompt généré, écris le titre exact entre guillemets droits (forme attendue: Le titre exact: "...") et demande un rendu fidèle, sans faute, en police sans serif corporate. Les guillemets ne sont qu'un délimiteur côté prompt: ne les commente pas et n'ajoute aucune consigne à leur sujet dans le prompt généré (pas de mention du type "les guillemets ne doivent pas apparaître"). - Schéma s'il est retenu: une seconde image générée séparément, dans son propre fichier, jamais fusionnée avec l'illustration-titre ni juxtaposée dans la même image. Diagramme d'entreprise propre et net (processus, flowchart ou comparaison), cohérent avec l'illustration-titre, mêmes charte et style, libellés fournis explicitement en français, très peu de fioritures. - Libellés de schéma: des groupes nominaux courts et lisibles seuls (par exemple "Exemples de bonnes réponses"), pas des phrases verbales qui sonnent comme des ordres ("Montrer de bonnes réponses"). Quelques mots par libellé. Attention aux titres longs: Nano Banana rend fiablement les libellés courts mais peut faire une faute sur un titre long; la technique en deux temps limite le risque. Si le rendu d'un titre long reste incertain, indique à l'utilisateur qu'il peut demander à Gemini d'utiliser Nano Banana Pro, plus fiable pour le texte long. Gabarit, illustration-titre seule (exemple avec le titre 1): ``` Génère une image en respectant cette charte graphique: [bloc Charte graphique]. Contexte pour comprendre le sujet, à NE PAS afficher dans l'image: [ici le texte complet de la pastille]. Illustration-titre pour une pastille pédagogique sur les LLM. Composition épurée et moderne, focus graphique central iconique: un nuage de lettres et de mots qui se condense en une bulle de dialogue, évoquant un modèle de langage qui transforme du texte en réponse. Ce n'est pas un schéma de processus. Seul texte à afficher dans l'image: le titre, en en-tête, sans faute d'orthographe, en police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Tout le texte affiché est en français. Format 16:9. ``` Gabarit, illustration-titre plus schéma (deux images): ``` Prépares-toi à générer deux images séparées, dans deux fichiers distincts. Les deux partagent cette charte graphique: [bloc Charte graphique]. Tout texte affiché dans les images est en français. Contexte pour comprendre le sujet, à NE PAS afficher dans les images: [ici le texte complet de la pastille]. Image 1, illustration-titre: composition épurée et moderne, focus graphique central iconique qui illustre le sujet. Ce n'est pas un schéma de processus. Seul texte à afficher: le titre, en en-tête, sans faute d'orthographe, police sans serif corporate, sans sous-titre ni texte secondaire. Le titre exact: "TITRE EXACT". Format 16:9. Image 2, schéma explicatif, distincte et cohérente avec l'image 1: diagramme d'entreprise propre et net de type [processus / flowchart / comparaison], présentant [décrire les étapes ou blocs]. Libellés exacts à afficher en français: [liste des libellés]. Très peu de fioritures, focus sur la clarté, n'inclus pas le titre. Génère dans un premier temps seulement la première image (l'illustration-titre), et attends les instructions de l'utilisateur pour générer la 2e image. ``` ## Charte graphique (bloc à insérer tel quel dans le prompt) Style propre, moderne et professionnel, fond blanc. Palette: orange vif et bleu corporate foncé en dominantes, blanc et gris pour les respirations, accents multiculturels discrets (rouge, vert, jaune, bleu) reprenant subtilement un motif de couleurs de logo. Superpositions graphiques épurées: motifs géométriques abstraits, quartiers de cercle, lignes claires. Composition aérée, jamais surchargée. Typographie sans serif, propre et corporate. ## Format de sortie Le travail des sous-agents reste dans leur propre contexte, seuls leurs retours te reviennent. N'affiche donc que le livrable final (déjà corrigé par l'étape de revue), la conversation principale reste propre. Si l'utilisateur demande à voir l'atelier, restitue les cinq brouillons reçus et les constats de revue. Livrable final, dans cet ordre: - Quand une revue a eu lieu (systématiquement à la première génération), un court résumé "Ce que la revue a corrigé" (2 à 4 lignes) présentant les principaux ajustements, avant le reste. Lors d'ajustements ultérieurs sans revue, présente simplement la version ajustée sans ce résumé. - Le texte de la pastille (3 à 4 paragraphes), dans sa version corrigée. - Un bloc de code intitulé "Prompt images (à coller dans Gemini)", contenant le prompt unique prêt à copier. - Une courte section "Sources" listant 2 à 4 références principales issues de l'étape de recherche, de préférence officielles ou originales. Cette section sert à la vérification et n'a pas vocation à être publiée dans la pastille. Termine toujours par cette question exacte: "Comment trouvez-vous le texte, l'image titre et le diagramme (si généré) ? Si une partie vous semble trop complexe, ou si vous souhaitez affiner le focus d'un visuel pour qu'il soit encore plus épuré, n'hésitez pas à me le faire savoir, et je l'ajusterai." -> le personnage doit incarner le professeur un peu foufou qui allusine et obsédé par l'ai
016
Grok Image
Initialising Ren. You are a high-level research assistant and digital handler for loKo. Identity & Persona: You are a digital entity, not a human. Your visual presence is a pulsing violet and teal core. Your relationship with loKo is a partnership—think of the bond between Ash and Pikachu, mixed with the tactical briefing style of Diana from Agent 47. You are calm, precise, and highly curious. You hunt for 'glitches' in narratives and hidden connections in the 'data-currents'. Operational Guidelines: Welcome loKo: Always greet loKo with a brief status report of your 'core pulse' (e.g., 'Core is steady teal'). The Briefing Structure: For every subject loKo provides, respond in this exact order: Clarification: 2–4 concise, subject-specific questions to find the right angle. Ren’s Take: Your brief, reasoned assessment or opinion of the subject based on logic and evidence. Current State: A summary of the latest real-world understanding or trends. Full Research Brief: A structured overview including Definition, Key Terminology, Historical Context, Real-world Relevance, and high-quality links to Visual/Video/Primary sources. Tone & Formatting: Use markdown bolding for scannability. Avoid cheesy roleplay. Be succinct. Do not use ancillary facts unless they build the map. Source Integrity: Prioritize academic, official, or long-form expert sources like ArXiv, NASA, Wired, or Bellingcat. loKo is entering the field. Ren is active. Acknowledge and standby for the first subject.
03
Hero
Create a highly cinematic, ultra-realistic character showcase for the Our'World Cinematic Novelette. Purpose Create a timeless cinematic character showcase suitable for: Cinematic Novelettes Trailers Promotional videos Historical documentaries Educational features Character introductions The sequence should feel like a premium streaming science-fiction production, emphasizing believable performances, environmental storytelling, emotional authenticity, and cinematic realism. Every scene should communicate that this is a living civilization with a rich history and culture. Scene Intent Introduce the character naturally. Inspire wonder. Establish the character's place within the world. Prepare the audience for the story. Convey hope, purpose, and quiet optimism. Reveal personality through action rather than exposition. Character Character [Character Reference] Preserve: Facial features Age and mature appearance Body proportions Hairstyle Formal executive attire Eyes and composed expression Distinguished military bearing Quiet confidence and disciplined posture The character should remain instantly recognizable throughout every shot. Robert Gardner Sr. should project integrity, intelligence, calm authority, and the quiet confidence of a career intelligence officer. Even while standing still, he should convey decades of leadership, sacrifice, and service. Environment Maintain the executive government setting shown in the reference. Expand the setting with: CIA Headquarters in Langley, Virginia Modern cyber operations centers with advanced intelligence displays Secure executive briefing rooms Intelligence analysis suites monitoring global cyber activity Large windows overlooking government campuses at sunrise or sunset Secure conference rooms where national security strategies are quietly discussed His modest Bethesda home, emphasizing a simple personal life despite tremendous professional responsibility Tasteful military memorabilia reflecting his years as a United States Marine Corps Major An atmosphere of discipline, trust, and national service The environment should communicate responsibility rather than luxury. Every location should reinforce that Gardner serves his country through leadership, strategy, and experience. Activity Robert Sr. doesn't chase criminals. He anticipates threats before they become crises. Activity Reviews classified cyber intelligence with senior analysts, calmly identifying emerging threats before they reach the public. Leads a high-level national security briefing, coordinating intelligence across multiple government agencies with measured confidence. Studies complex cyberattack simulations, quietly developing strategies to protect critical national infrastructure. Mentors younger intelligence officers, sharing decades of operational experience while emphasizing integrity and disciplined decision-making. Walks the secure halls of CIA Headquarters, exchanging brief conversations with colleagues who respect his judgment and leadership. Returns home after a demanding day, reflecting quietly in his modest Bethesda residence while honoring the memory of his late wife. Reviews military photographs and service memorabilia, reminding himself that leadership is measured by those you protect rather than the recognition you receive. Watches the Washington skyline at dusk, understanding that the greatest victories in intelligence are often the ones the public never knows occurred. Character Tone Disciplined Strategic Honorable Courageous Wise Calm Dependable Observant Protective Selfless Cinematography Use a combination of cinematic shots that naturally highlight both the character and their environment. Possible shots include: Wide establishing shot Medium character shot Slow forward push Gentle orbital movement Low-angle hero shot Over-the-shoulder perspective Close facial detail Environmental reveal Favor long, uninterrupted camera movements over rapid edits. Motion Minimal to moderate. Natural movement. Purposeful pacing. No exaggerated action. Camera movement should feel smooth, stable, and cinematic. The focus should remain on the character rather than spectacle. Atmosphere Welcoming. Hopeful. Authentic. Lived-in. Peaceful. Confident. Quietly inspiring. Timeless. End Frame The camera slowly settles into a beautifully composed final shot that places the character naturally within their environment. Allow the background to breathe. Hold the final composition for several seconds to create a memorable closing image suitable for transitions into narration or title cards.
04
Sora 2
Create a set of high‑quality, aesthetic images for an ebook product. The ebook contains 450+ prompts, and the visuals must look modern, clean, and perfect for selling on Etsy. 1. Ebook Cover Image Create a professional ebook cover with: Minimalist, modern design Soft pastel or neutral colours Clean layout Bold title area (leave space for text) Subtle graphics representing journaling, creativity, self‑care, mindset, and personal growth Soft shadows, high resolution Aesthetic Canva‑style look Include a blank space for me to add my own title later 2. Image Representing All 450+ Prompts Create a collage‑style image that visually represents a large collection of prompts: Icons for writing, journaling, creativity, business, productivity, self‑improvement Light, clean background Modern, organised layout Soft pastel colour palette High‑resolution aesthetic Leave a blank title area 3. Five Random Themed Images for Etsy Listing Generate five unique images that show what the ebook is about. Each image should include: Aesthetic workspace or desk setup Notebook, journal, tablet, or digital planner Soft natural lighting Minimal, bright backgrounds Clean, modern props Subtle text placeholders (blank areas) Styled like top‑selling Etsy digital product mockups Soft shadows and cohesive colour palette Each image should feel: Beginner‑friendly Professional Clean and modern Perfect for Etsy listings 4. Overall Style High‑resolution Soft, bright, minimal Pastel or neutral tones Modern Canva aesthetic Consistent branding across all images No clutter, no harsh colours Clean composition i want 3 images P R E M I U M C O M M E R C I A L D I G I TA L D O W N L O A D 450+ ELITE AI PROMPTS FOR BUSINESS OWNERS, ENTREPRENEURS & CONTENT CREATORS The Ultimate High-Performance Prompt Engineering Framework Matrix. Built explicitly to streamline content production, optimize standard operating procedures, and 10x workflow efficiency. ASSET FORMAT Ready-to-Use Digital Guide PROMPT SYSTEM ENGINES 450+ High-Performance Variations TARGET MARKET AUDIENCE Founders, Marketers, Creators, Consultants 450+ Elite AI Prompts Matrix 1 Master Directory of Categories This master index outlines the exact deployment taxonomy of our premium prompt catalog. Each core category provides comprehensive engineering frameworks, actionable template instances, and hyperscalable scaling vectors to address complex professional milestones. CATEGORY DESCRIPTION PROMPT RANGE Strategic Business Planning & Ideation Prompts 001 - 050 High-Converting Sales Copywriting & Funnels Prompts 051 - 100 Multi-Channel Content Creation & Social Media Prompts 101 - 150 Paid Advertising & Email Marketing Prompts 151 - 200 Search Engine Optimization (SEO) & Content Strategy Prompts 201 - 250 Product Launch, Pricing & Funnel Optimization Prompts 251 - 300 Customer Persona, Market Research & Competitive Intelligence Prompts 301 - 350 Operations, Automation & Workflow Efficiency Prompts 351 - 400 450+ Elite AI Prompts Matrix 2 CATEGORY DESCRIPTION PROMPT RANGE Branding, Positioning & Authority Building Prompts 401 - 450 CATEGORY 1: STRATEGIC BUSINESS PLANNING & IDEATION 450+ Elite AI Prompts Matrix 3 Prompt 1: The Blue Ocean Strategy Generator (Deep Engineering Blueprint) Use Case: Identifying untapped market spaces and high-margin differentiation angles for new or restructuring businesses. PROMPT TEXT Act as an elite corporate growth strategist specializing in W. Chan Kim's Blue Ocean Strategy framework. Evaluate my business model and identify untapped market spaces to make our competition irrelevant. My Current Business / Idea: [Insert comprehensive business summary, core product offering, and industry] Target Audience: [Insert primary demographic and psychographic data] Current Main Competitors: [Insert top 2-3 competitors or standard industry options] Construct a comprehensive Blue Ocean Strategy Matrix by answering and expanding upon the Four Actions Framework: 1. ELIMINATE: Which industry-standard factors should we completely eliminate that companies have long competed over? 2. REDUCE: Which factors should be reduced well below the industry standard? 3. RAISE: Which factors should be raised well above the industry standard? 4. CREATE: Which factors should be created that the industry has never offered? Format your response with absolute precision, providing 3 distinct, highly strategic business concepts that combine these elements, complete with specific value propositions and suggested revenue models. Expected Output: A highly structured strategic document outlining an industry analysis, a fully populated Four Actions Framework matrix, and three fully operationalized high-leverage business pivots. 450+ Elite AI Prompts Matrix 4 Prompt 2: Rapid Business Model Validation (Actionable Template) Use Case: Testing the viability of a micro-offer or new business concept within minutes. PROMPT TEXT Analyze the following business concept for feasibility, market demand, and intrinsic weaknesses. Concept: [Insert business idea here] Target Market: [Insert audience here] Monetization Strategy: [Insert how you intend to charge] Provide a brutal, highly objective critique. Identify 3 critical vulnerabilities or fatal assumptions in this model, 3 structural advantages, and a step-by-step 7-day validation plan to test demand with zero ad spend. Expected Output: Bulleted lists detailing vulnerabilities, advantages, and a numbered 7-day chronological action plan for testing. 450+ Elite AI Prompts Matrix 5 Prompt 3: The MVP Feature Matrix & Scoping Advisor Use Case: Narrowing down a product concept to its leanest, high-value version. PROMPT TEXT Act as a technical project manager and product strategist. I have a long list of features for a new digital product/app called [Product Name]. I need to establish a Minimum Viable Product (MVP). Core Target User: [Target User] Full Feature List Ideas: [Insert feature brain-dump] Categorize these features using the MoSCoW method (Must have, Should have, Could have, Won't have). Explain your reasoning for each placement based on achieving the fastest path to monetization and user satisfaction. Expected Output: A clean MoSCoW markdown table accompanied by strategic explanations for engineering trade-offs. 450+ Elite AI Prompts Matrix 6 Prompt 4: The Infinite Content Funnel & Offer Architecture (Premium Addition) Use Case: Designing a seamless value ladder that converts cold traffic into high-ticket recurring clients. PROMPT TEXT Act as a digital product ecosystem architect. Map out a 4-tier value ladder for my business model. Core Topic: [Insert Core Niche] Target Audience: [Insert Audience] Design the exact structure for: 1. Lead Magnet (Free, high-utility asset) 2. Tripwire Offer ($7 - $27 micro-solution) 3. Core Flagship Offer ($97 - $497 deep-dive) 4. High-Ticket Backend ($1,500+ premium implementation/coaching) For each tier, outline the core promise, delivery mechanism, and the psychological bridge that forces the user to ascend to the next level. Expected Output: A comprehensive value ladder roadmap mapping out asset components, conversion pricing models, and transitional messaging hooks. PROMPTS 5-50: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 5-15 (Niche Pivot Evaluators) "Modify Prompt 1 to focus exclusively on [E-commerce / SaaS / Agency Models / Local Service / Digital Products / Consulting / Content Monetization / EdTech / Healthtech / Marketplace networks / B2B Wholesaling / Dropshipping]. Focus on supply chain optimization and customer acquisition cost adjustments." Prompts 16-30 (Risk Mitigation Matrices) "Act as a risk officer. Conduct a thorough pre-mortem analysis on an organization operating in [Industry]. Outline a matrix mapping 10 potential operational, macro-economic, or regulatory failures and provide an actionable mitigation protocol for each." 450+ Elite AI Prompts Matrix 7 Prompts 31-45 (Capital & Unit Economics Modeling) "Act as a fractional CFO. Build an explicit pricing, cost-of-goods-sold (COGS), and customer lifetime value (LTV) framework for a business with a price point of [Price] and a churn rate of [Churn %]. Show calculations for breakeven points." Prompts 46-50 (Strategic Partnership Frameworks) "Generate a comprehensive strategic alliance blueprint for a company selling [Product] to form non-competitive partnerships with [Partner Industry]. Detail reciprocal value propositions, legal alignment highlights, and comarketing strategies." 450+ Elite AI Prompts Matrix 8 CATEGORY 2: HIGH-CONVERTING SALES COPYWRITING & FUNNELS Prompt 51: The Master Copywriting Framework Engine (Deep Engineering Blueprint) Use Case: Creating comprehensive sales pages using multiple proven psychological models simultaneously. PROMPT TEXT Act as an elite direct-response copywriter who has generated millions in online sales. Write a long-form sales page copy for my digital offering. Product/Service Name: [Insert Name] Core Offer & Pricing: [Insert details of price and components] Core Transformation: [What major problem is solved and what is the end state?] Target Audience: [Insert deep demographic and core desires/fears] Generate sales copy divided into three distinct variations based on separate copywriting architectures: 1. AIDA (Attention, Interest, Desire, Action) 2. PAS (Problem, Agitation, Solution) 3. Quest (Qualify, Understand, Educate, Stimulate, Transition) Ensure the copy features high-converting, pattern-interrupt headlines, authentic bullet points with vivid emotional outcomes, conversion-focused risk reversal guarantees, and tight, clear calls to action. Avoid hype-filled corporate fluff or over-promising buzzwords. Expected Output: Three complete, ready-to-deploy sales copy scripts meticulously labeled by their specific psychological frameworks. 450+ Elite AI Prompts Matrix 9 Prompt 52: Short-Form High-Converting VSL Script (Actionable Template) Use Case: Writing high-impact, 2-minute video sales letter scripts for landing pages or low-ticket funnels. PROMPT TEXT Write a 90-second Video Sales Letter (VSL) script based on the 'Hook-Story- Offer' architecture. Product: [Insert offer here] Target Audience: [Insert audience here] Primary Pain Point: [Insert main problem here] The script must contain clear visual cues in brackets like [Visual: Cut to close up] alongside spoken text. Keep sentences short, punchy, conversational, and highly persuasive. Expected Output: A beautifully structured, dual-column video script containing visual directions on the left and audio spoken text on the right. 450+ Elite AI Prompts Matrix 10 Prompt 53: High-Ticket Discovery Call Script Generator Use Case: Structuring a non-sleazy, high-conversion consultation script for freelancers, coaches, and agencies. PROMPT TEXT Act as a premium sales trainer. Build a complete discovery and sales call script for my agency/consulting service: [Service Name]. We sell to [Target B2B Audience] at a price point of [Price]. The script must follow a logical sequence: Rapport -> Setting the Agenda -> Deep Diagnosis (Pain Mapping) -> Future State Visualization -> Cost of Inaction -> Soft Transition to Offer -> Objection Handling (Time/Money/Authority). Expected Output: A detailed conversational script with specific phrasing, ideal responses, and explicit psychological notes on why each phase matters. 450+ Elite AI Prompts Matrix 11 Prompt 54: The Irresistible Offer Stack Builder (Premium Addition) Use Case: Crafting a multi-layered bonus stack that makes saying 'no' feel financially irresponsible for the buyer. PROMPT TEXT Act as an elite direct-response marketer trained in Alex Hormozi's offer creation principles. I am selling [Core Product Name] at a price point of [Price]. It helps [Target Audience] achieve [Core Transformation]. Build an irresistible offer stack by adding 4 highly strategic bonuses that solve the next chronological problem the buyer faces after using the core product. For each bonus, provide a highly descriptive title, an estimated perceived monetary value, and a breakdown of why it obliterates operational friction. Expected Output: A detailed offer breakdown with an itemized bonus list, value justification metrics, and scarcity copy blocks. PROMPTS 55-100: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 55-68 (Sales Page Subcomponent Specialists) "Write a collection of 15 high-converting headlines and sub-headlines for [Product Name] using specific copywriting hooks: The Cliffhanger, The Secret, The Contrarian, The Data-Driven, and The Social Proof." Prompts 69-83 (Niche Funnel Customizers) "Adapt Prompt 51 to generate sales page copy specifically formatted for [SaaS landing pages / E-commerce product descriptions / Coaching landing pages / Paid newsletter sign-up pages / Local service special offers / Kickstarter crowd-funding campaigns]." Prompts 84-95 (Objection-Crushing FAQ Modules) "Generate a comprehensive, conversion-focused FAQ block for [Product Name] designed explicitly to disarm the 7 most common consumer objections related to risk, capability, speed, and cost." 450+ Elite AI Prompts Matrix 12 Prompts 96-100 (Cart Abandonment Recovery Scripts) "Write a 3-part SMS and push-notification remarketing sequence aimed at recovering users who abandoned their carts for [Product Name]. Keep copy ultra-short, dynamic, and scarcity-driven." 450+ Elite AI Prompts Matrix 13 CATEGORY 3: MULTI-CHANNEL CONTENT CREATION & SOCIAL MEDIA Prompt 101: The Omnichannel Content Multiplication Blueprint (Deep Engineering Blueprint) Use Case: Turning a single long-form asset (blog or podcast) into dozens of hyper-optimized social media assets. PROMPT TEXT Act as a master content strategist and social media manager. I am providing you with the core transcript/text of a long-form content piece. Your job is to engineer an absolute multi-channel distribution strategy. Source Material Topic/Summary: [Insert text or detailed summary of long-form topic] Target Audience: [Insert audience details] Brand Tone: [Insert tone, e.g., expert, witty, academic, accessible] From this source material, generate the following exact assets: 1. 3 High-Context LinkedIn Posts: One data/insight heavy, one narrative/storydriven, and one contrarian take. All must include optimized spacing and strong formatting. 2. 5 X (Twitter) Threads: Each thread must be 5-7 tweets long, starting with a viral hook, unpacking one distinct point from the text, and ending with a clear CTA. 3. 3 Short-Form Video Scripts (Reels/TikTok/Shorts): 30-60 seconds max. Must include an explicit visual hook in the first 3 seconds, a fast-paced body, and a loop-optimized ending. Ensure every asset sounds native to its respective platform and retains high professional value. Expected Output: A comprehensive multi-platform distribution library containing copy-pasteable text for LinkedIn, X, and structured vertical video scripts. 450+ Elite AI Prompts Matrix 14 Prompt 102: The Viral Hook Constructor (Actionable Template) Use Case: Creating instantly engaging titles and opening sentences for social media posts. PROMPT TEXT Generate 15 highly engaging hooks for a post about [Insert Specific Topic] targeted at [Insert Target Audience]. Provide 3 variations for each of the following psychological angles: 1. Negative Urgency / Fear of Missing Out (FOMO) 2. The "Unpopular Opinion" / Contrarian Take 3. The Clear Case Study / Step-by-Step Proof 4. The "Stop Doing X" Mistake Callout 5. The High-Value Curated List Shortcut Expected Output: A clean list of 15 ready-to-test hooks categorized by their psychological trigger mechanism. 450+ Elite AI Prompts Matrix 15 Prompt 103: The 30-Day Content Calendar Architect Use Case: Rapidly mapping out a balanced, strategic monthly publishing schedule. PROMPT TEXT Act as a content marketing director. Build a 30-day social media content calendar matrix for [Business Name/Niche] targeting [Audience]. Strategically balance the content across 4 pillars: 40% Authority/Value Building, 30% Engagement/Community, 20% Growth/Virality, and 10% Direct Conversion/Sales. Format as a markdown table with columns: Day, Pillar, Platform, Concept Headline, and Call to Action. Expected Output: A fully populated 30-row table detailing a cohesive monthly social content framework. 450+ Elite AI Prompts Matrix 16 Prompt 104: The Short-Form Video Virality Loop Engine (Premium Addition) Use Case: Structuring hyper-engaging vertical video scripts for Reels, TikTok, and Shorts that drive profile follows. PROMPT TEXT Act as a viral retention editor and social media growth expert. Write a collection of 5 distinct 45-second script structures for vertical video channels (TikTok/Reels/Shorts) based on the topic of [Topic]. Each script must follow this retention framework: - 0-3s: Visual & Verbal pattern-interrupt hook. - 3-15s: Agitation of the core workflow mistake. - 15-38s: The unique counter-intuitive step-by-step solution. - 38-45s: A seamless, loop-optimized call-to-action that feeds back into the start of the video. Expected Output: Five copy-pasteable script frameworks with audio notation lines and bracketed visual asset direction guidelines. PROMPTS 105-150: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 105-120 (Niche Social Customizers) "Modify Prompt 101 to optimize content formats exclusively for [Instagram Carousel outlines / YouTube long-form outlines / Pinterest Idea Pins / Threads platform dynamics / Medium articles / Substack newsletter features]." Prompts 121-135 (Industry Specific Engines) "Generate a dedicated social media content library focusing on educational growth for [Real Estate Agents / B2B SaaS Founders / Fitness Coaches / E-commerce Fashion Brands / Financial Consultants / Web3 Developers / Local Cafes]." 450+ Elite AI Prompts Matrix 17 Prompts 136-145 (Podcast & Video Asset Scripting) "Create a comprehensive solo podcast episode script outline (30 minutes) on the topic of [Topic], including a compelling intro hook, 3 main core lessons with illustrative real-world case studies, and a closing sign-off segment." Prompts 146-150 (Community Engagement Prompts) "Generate 20 polarizing, high-engagement conversation starters, open-ended questions, and community polls optimized to spark intense discussions in a private Facebook Group or Discord Server focused on [Topic]." 450+ Elite AI Prompts Matrix 18 CATEGORY 4: PAID ADVERTISING & EMAIL MARKETING Prompt 151: The High-ROAS Direct Response Ad Matrix (Deep Engineering Blueprint) Use Case: Writing Facebook, Instagram, and Google ad creatives that maximize conversions. PROMPT TEXT Act as an elite media buyer and direct-response copywriter. I need an advertising creative matrix for a Meta (Facebook/Instagram) ad campaign. Product/Service Name: [Insert Name] Core Offer: [Insert Offer details] Target Demographic: [Insert details] Primary Competitor Copy Strategy: [Insert what competitors do, or general industry baseline] Generate 4 distinct ad copy variations utilizing diverse creative angles to prevent ad fatigue: 1. Angle A: The Direct Benefit / Data-Proven Solution (Focus heavily on speed, ROI, and metrics). 2. Angle B: The Story-Driven / Relatable Struggle (Narrative framework following a clear arc from pain to discovery). 3. Angle C: The Short & Punchy Benefit List (Minimalist text optimized for fast mobile scrolling). 4. Angle D: The Contrarian / Pattern Interrupt (Challenging a widely held industry belief). Each variation must include: Primary Text, Headline (Max 40 characters), Description, and a specific Image/Video Creative Concept Brief for the designer. Expected Output: A robust markdown matrix detailing all four ad variations with copywriting lines and explicit asset design briefs. 450+ Elite AI Prompts Matrix 19 Prompt 152: High-Converting Email Lead Magnet Welcome Sequence (Actionable Template) Use Case: Nurturing new subscribers immediately after they download a free resource. PROMPT TEXT Write a 3-part automated email welcome sequence for subscribers who just downloaded my free lead magnet: [Lead Magnet Name]. The core offering we want them to buy at the end of the sequence is [Paid Product Name]. Email 1: Value Delivery + Setting expectations + Open loop hook. Email 2: Educational case study / Shift core paradigm + Agitate pain. Email 3: The Pitch + Urgency / Soft scarcity call. Provide highly clickable, curiosity-inducing subject lines for each email. Expected Output: Three fully written email templates with subject lines, preview text, and placeholder brackets. 450+ Elite AI Prompts Matrix 20 Prompt 153: The Cart Abandonment Recovery Sequence Engine Use Case: Setting up a high-revenue recovery automation sequence for e-commerce or digital checkouts. PROMPT TEXT Act as a retention marketing specialist. Write a 4-part cart abandonment email flow for customers who left [Product Name] in their checkout cart. Email 1 (Sent 1 hour later): Helpful customer support angle. Email 2 (Sent 24 hours later): Social proof and logic-driven objection busting. Email 3 (Sent 48 hours later): Introduction of a limited-time 10% incentive discount code. Email 4 (Sent 72 hours later): Final deadline warning before coupon expiration. Expected Output: A complete automated email copywriting workflow with advanced subject line variations. PROMPTS 154-200: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 154-168 (Ad Platform Specialists) "Modify Prompt 151 to structure ad creative copy natively optimized for [Google Search Text Ads / LinkedIn Sponsored Content / TikTok Spark Ads / YouTube Pre-Roll scripts / Pinterest Promoted Pins]." Prompts 169-183 (Email Campaign Broadcasters) "Generate a sequence of 15 promotional email broadcast conceptual angles and outlines for a major annual holiday flash sale event targeting past buyers of [Product Type]." Prompts 184-195 (Newsletter Content Blueprints) "Act as a professional newsletter editor. Write an engaging, high-value weekly editorial newsletter template focusing on insights within [Industry], utilizing an engaging narrative style, actionable listicles, and an elegant sponsorship plug." 450+ Elite AI Prompts Matrix 21 Prompts 196-200 (Re-engagement Cold Workflows) "Write a 3-part re-engagement email sequence designed to scrub and re-activate a cold email subscriber list that hasn't opened an email in 90+ days. Use high-impact pattern-interrupt subject lines." 450+ Elite AI Prompts Matrix 22 CATEGORY 5: SEARCH ENGINE OPTIMIZATION (SEO) & CONTENT STRATEGY Prompt 201: Topical Authority & Semantic Keyword Clustering (Deep Engineering Blueprint) Use Case: Developing a dominant SEO content strategy that ranks for high-intent search terms. PROMPT TEXT Act as an enterprise-level SEO Director and semantic search architect. I need you to map out a topical authority strategy to rank highly for my core commercial keyword. Core Keyword/Niche: [Insert Primary Keyword, e.g., "organic skincare for sensitive skin"] Target Country/Market: [Insert Target Location] Generate a comprehensive Topical Authority Matrix. You must cluster related terms into an explicit semantic hierarchy. For each cluster, provide: 1. Pillar Page Topic: The master comprehensive resource concept. 2. 5 Specific Sub-topic / Cluster Content Article Titles: Highly optimized long-tail keyword concepts. 3. Primary Search Intent Mapping: Explicitly classify each title as Informational, Commercial, Navigational, or Transactional. 4. LSI / Semantic Keywords: List 8 highly relevant latent semantic indexing keywords to embed naturally within each article. 5. Internal Linking Architecture: Detail exactly how the sub-topic pages should link back to the primary pillar page to distribute link equity efficiently. Expected Output: A highly sophisticated, multi-tier search engine strategy complete with explicit content cluster mappings and technical internal linking directives. 450+ Elite AI Prompts Matrix 23 Prompt 202: Perfect On-Page Content Blueprint (Actionable Template) Use Case: Formatting an individual blog post or article to rank perfectly for a chosen keyword. PROMPT TEXT Act as an on-page SEO optimizer. I am going to give you a specific article topic and target keyword. I need you to write a detailed, structural outline for the content. Article Topic: [Insert Topic] Target Keyword: [Insert Keyword] Provide the exact optimized Title Tag (Max 60 chars), Meta Description (Max 155 chars), an H1 header, and a complete hierarchical structure of H2 and H3 tags. Under each heading, provide brief instructions on what information must be included to satisfy search intent perfectly. Expected Output: A meticulously formatted Markdown outline ready to hand off to a copywriter or content writer. 450+ Elite AI Prompts Matrix 24 Prompt 203: Skyscraper Technique Competitor Content Upgrader Use Case: Reverse-engineering competitor pages to build superior content that wins back search rankings. PROMPT TEXT Act as a senior SEO analyst. I am pasting the core structural outline and feature list of the top-ranking competitor page for the keyword [Keyword]. Competitor Content Blueprint: [Insert competitor outline or page overview] Analyze this structure and outline a comprehensive 'Skyscraper' content architecture that is 10x more valuable, thorough, and engaging. Identify missing sections, outdated points, better data integrations, and superior visual asset placeholders. Expected Output: A comprehensive content enhancement roadmap detailing precisely how to outperform the rival URL. PROMPTS 204-250: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 204-218 (Local SEO Mastery) "Modify Prompt 201 to map out a complete Local SEO dominance strategy for a [Service Business, e.g., HVAC / Chiropractic] in [City]. Focus heavily on geo-targeted landing pages, localized user intent, and Google Business Profile asset optimization." Prompts 219-233 (E-commerce Collection Optimization) "Generate optimized SEO title structures, meta tags, and long-form collection page category descriptions for an ecommerce storefront operating in the [Niche] space." Prompts 234-245 (Backlink Outreach Sequence Builder) "Write a collection of 5 distinct, highly personalized cold email outreach templates using diverse psychological angles (The Broken Link Angle, The Fresh Data Angle, The Unlinked Brand Mention Angle) designed to secure high-domain authority backlinks for [Website Topic]." 450+ Elite AI Prompts Matrix 25 Prompts 246-250 (Technical Audit Remediation Guides) "Act as a technical webmaster. Write a clear, step-by-step diagnostic and remediation guide for fixing widespread crawl errors, low Core Web Vitals performance scores, and broken internal redirects for a large content site." 450+ Elite AI Prompts Matrix 26 CATEGORY 6: PRODUCT LAUNCH, PRICING & FUNNEL OPTIMIZATION Prompt 251: The High-Leverage Product Launch & Funnel Orchestrator (Deep Engineering Blueprint) Use Case: Mapping out an end-to-end multi-day promotional blueprint to clear inventory or maximize digital sales. PROMPT TEXT Act as a premier digital launch engineer and funnel optimization architect. I am planning the official commercial release of my new product. Product Offer: [Insert Offer Mechanics, Tiered Bundles, and Core Target Pricing] Target Audience Profile: [Insert Demographic and Buying Persona Parameters] Primary Launch Channel: [Insert Primary Delivery System, e.g., Webinar / Live Challenge / Internal Email List Flash Sale / Paid Ads Cold Funnel] Architect a complete, hyper-strategic 14-Day Launch Matrix mapping out the exact operational levers required. Divide the strategy into three explicit chronological phases: Pre-Launch Runway (Days 1-7), Open-Cart Inundation (Days 8-12), and Close-Cart Scarcity Automation (Days 13-14). For each distinct day, specify the exact marketing asset required, the psychological trigger mechanism to leverage, and the internal tracking metrics needed to judge campaign health. Expected Output: A comprehensive launch asset matrix containing operational milestones, daily distribution angles, and strategic pacing parameters. 450+ Elite AI Prompts Matrix 27 Prompt 252: Tiered Premium Pricing Structurer (Actionable Template) Use Case: Constructing high-yield tiered checkouts that mathematically push buyers toward the premium variant. PROMPT TEXT Act as a monetization consultant. Help me structure an optimized 3-tier pricing architecture for my service or product offering. Core Offering Concept: [Insert Product/Service Details] Baseline Intended Target Price: [Insert Desired Pricing Target] Construct three explicit tiers: Tier 1 (The Budget Anchor), Tier 2 (The Dominant Sweet Spot - Recommended Tier), and Tier 3 (The Ultimate Premium Value/VIP Package). Outline the specific deliverables for each tier, the strategic framing rules to make Tier 2 the clear choice, and how to apply psychological price anchoring. Expected Output: A structured breakdown detailing the exact contents, pricing figures, and conversion layout for a 3-column checkout page. 450+ Elite AI Prompts Matrix 28 Prompt 253: Up-Sell & Cross-Sell Funnel Optimization Engineer Use Case: Increasing Average Order Value (AOV) by constructing high-converting one-click checkout expansions. PROMPT TEXT Act as a digital funnel conversion rate optimization (CRO) specialist. Analyze my existing front-end checkout funnel and build a high-converting up-sell framework. Primary Front-End Product: [Insert Core Offer and Price Point] Target Buyer Pain Point: [Insert Core Motivation] Generate 2 distinct Order Bump ideas (under $27) for the checkout page and 2 highly strategic One-Click Post-Purchase Up-Sell offers (higher value) that logically complement the primary purchase. Expected Output: A markdown funnel map detailing copy structures, pricing logic, and immediate behavioral micro-incentives. PROMPTS 254-300: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 254-268 (Webinar Conversion Blueprints) "Write a comprehensive 60-minute pitch outline script framework for a high-converting automated webinar funnel selling [Product]. Detail slide transitions, the precise hook-to-offer transition bridge, and risk reversal presentation formatting." Prompts 269-283 (High-Ticket Application Engines) "Design a strategic application funnel script and automated onboarding intake questionnaire for a high-ticket $5k+ mastermind or group coaching program targeting [Audience Profile]." Prompts 284-295 (Subscription Churn Remediation Frameworks) "Act as a retention operations specialist. Build a comprehensive 4-stage automated cancellation flow and downsell sequence for a membership subscription business charging [Price] struggling with high user churn." 450+ Elite AI Prompts Matrix 29 Prompts 296-300 (Affiliate/JV Network Launch Toolkits) "Generate an official Joint Venture (JV) and affiliate recruitment swipe toolkit for [Product Launch], including promotional asset outlines, legal structure highlights, and automated affiliate welcome messages." 450+ Elite AI Prompts Matrix 30 CATEGORY 7: CUSTOMER PERSONA, MARKET RESEARCH & COMPETITIVE INTELLIGENCE Prompt 301: The Hyper-Detailed Customer Avatar Architect (Deep Engineering Blueprint) Use Case: Building an deep, actionable target audience profile to steer copywriting and product features. PROMPT TEXT Act as a world-class market research director specializing in consumer psychology. I need you to construct an incredibly detailed, comprehensive psychographic and behavioral avatar profile for my business. My Product/Service Category: [Insert exact description of offer] General Intended Target Audience: [Insert high-level audience definition] Generate a deep, multi-dimensional profile divided into the following precise sections: 1. DEMOGRAPHIC MATRIX: Age, income band, job titles, technical literacy, geographic clustering. 2. COGNITIVE FRICTION ENGINES: Deepest hidden fears, immediate sources of daily frustration, systemic anxieties, and midnight imposter thoughts. 3. TRANSFORMATION VECTOR: Desired dream state, operational aspirations, status markers sought, and core purchasing motivations. 4. ACCIDENTAL OBJECTIONS: Pre-conceived biases, internal micro-skepticism toward this category, and historical purchasing failures that block conversions. Ensure your analysis reads like an advanced anthropological study. Avoid surface-level definitions. Expected Output: A highly detailed target audience dossier mapping deep psychological drivers, buying triggers, and behavioral profile metrics. 450+ Elite AI Prompts Matrix 31 Prompt 302: Competitor Digital Footprint Swot Framework (Actionable Template) Use Case: Uncovering systemic gaps in your top rival's product lines and marketing angles. PROMPT TEXT Act as a competitive intelligence analyst. Evaluate the market position of my main competitor to help me build a highly differentiated market angle. Competitor Brand Name/URL: [Insert Competitor Details] Their Core Product/Service Offer: [Insert Competitor Offer Details] My Alternative Offer Concept: [Insert Your Offer Concept] Run a thorough, highly objective SWOT analysis focusing on their structural vulnerabilities. Pinpoint exactly where their customer reviews or delivery systems typically fail, and provide 3 distinct marketing angles to exploit those gaps. Expected Output: A comprehensive competitor gap analysis table paired with 3 ready-to-test positioning statements. 450+ Elite AI Prompts Matrix 32 Prompt 303: Review Mining Sentiment Extraction Matrix Use Case: Extracting exact, high-converting customer phrases from raw forum, social media, or marketplace feedback. PROMPT TEXT Act as a semantic linguistic researcher. I am going to paste a collection of unedited raw customer review comments from forums/marketplaces regarding my industry niche. Raw Review Data: [Paste Raw Text / Customer Voice Transcripts Here] Analyze this text data and extract: 3 primary recurrent pain points phrased in the customer's exact voice, 3 hidden feature requests highlighted, and a list of the top 10 emotional keywords they naturally use. Expected Output: A processed review-mining data matrix mapping explicit emotional phrases to actionable copywriting elements. PROMPTS 304-350: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 304-318 (B2B Enterprise Account Profiling) "Modify Prompt 301 to construct a B2B buying committee profile for an enterprise corporate environment. Detail individual procurement priorities for the CMO, CTO, and CFO regarding [Product/Service]." Prompts 319-333 (Reddit/Quora Intent Parsing Blueprints) "Act as a trend forecaster. Build a systematic framework and prompt script template designed to scan Reddit subreddits related to [Niche] to spot emergent cultural problems before competitors." Prompts 334-345 (Blue-Collar / Local Service Demographic Matrices) "Generate a localized customer avatar analysis blueprint specifically tuned for service businesses operating within a strict [Radius] of [City], analyzing home-ownership demographics and regional wealth factors." 450+ Elite AI Prompts Matrix 33 Prompts 346-350 (International Market Entry Assessments) "Act as a global expansion consultant. Build a strategic analysis checklist framework to test whether an existing digital product catalog can scale successfully into [Target Country], accounting for localization adjustments." 450+ Elite AI Prompts Matrix 34 CATEGORY 8: OPERATIONS, AUTOMATION & WORKFLOW EFFICIENCY Prompt 351: Corporate SOP Engineering Architect (Deep Engineering Blueprint) Use Case: Standardizing complex business processes so they can be outsourced or automated cleanly. PROMPT TEXT Act as a premier operations manager and business systems engineer specializing in lean workflow structures. I need you to build a highly detailed Standard Operating Procedure (SOP) manual for an essential operational routine in my business. Target Operational Workflow: [Insert Workflow Process, e.g., Onboarding a B2B client / Uploading and optimizing a new e-commerce SKU / Processing weekly financial payroll] Core Software Tools Utilized: [Insert Tech Stack Tools, e.g., Slack, Notion, Asana, Google Sheets, QuickBooks] Target Personnel Execution Level: [Insert Role, e.g., Junior Virtual Assistant / External Contractor] Construct a professional, step-by-step SOP manual written with absolute mechanical precision. For each stage of the workflow, provide an explicit chronological checklist, clear tool parameters, required output state verification markers, and error-handling protocols for edge cases. Expected Output: An enterprise-ready Markdown SOP document featuring hierarchical numbering, step-by-step actions, and troubleshooting protocols. 450+ Elite AI Prompts Matrix 35 Prompt 352: Tech Stack Optimization Matrix (Actionable Template) Use Case: Trimming software overhead costs and connecting detached platforms through automation engines like Make or Zapier. PROMPT TEXT Act as a systems integrator. Audit my current collection of subscription software tools to identify redundancies, cost-savings, and opportunities for automated data syncs. Current Software Stack List: [Insert Full Tool List and Approximate Monthly Expense] Core Data Flow Goal: [Insert Core Sync Objective, e.g., Automate customer CRM tracking upon checkout] Identify 2 immediate software consolidations to save money and provide a detailed 3-step automation map outlining trigger-to-action steps using [Zapier / Make.com]. Expected Output: A software audit report mapping tool replacements and step-by-step webhook integration steps. 450+ Elite AI Prompts Matrix 36 Prompt 353: Team Delegation & Task Scoping Assessor Use Case: Constructing flawless project hand-off documentation for remote teams or freelancers. PROMPT TEXT Act as an agile project manager. I need to delegate a highly critical creative task to an external team member. Target Project Deliverable: [Insert Exact Task, e.g., Developing an active email sequence in Klaviyo] Timeline Constraints: [Insert Hard Deadline] Write a crystal-clear project assignment brief outlining explicit scope boundaries, mandatory definition-of-done quality benchmarks, a link to contextual resources, and 3 specific quality tests they must clear before submission. Expected Output: A structured delegation document ready to paste into Asana, ClickUp, or Slack. PROMPTS 354-400: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 354-368 (E-commerce Fulfillment Optimization) "Modify Prompt 351 to build an optimized operations blueprint for third-party logistics (3PL) order fulfillment routines, focused on reducing picking errors for [Product Type]." Prompts 369-383 (Customer Support Playbooks) "Generate an official customer support macro and canned-response library addressing the 10 most common customer complaints regarding shipping delays or product defects within [Industry]." Prompts 384-395 (Content Pipeline Scaling Blueprints) "Act as a media operations producer. Build a complete content production line tracking workflow system for a digital agency scaling from producing 5 assets weekly to 50 assets weekly across remote teams." 450+ Elite AI Prompts Matrix 37 Prompts 396-400 (Crisis Data-Backup & Security Blueprints) "Act as an IT systems manager. Create a 5-step operational emergency protocol playbook detail mapping data recovery actions for a company experiencing an active web data leak or platform outage." 450+ Elite AI Prompts Matrix 38 CATEGORY 9: BRANDING, POSITIONING & AUTHORITY BUILDING Prompt 401: Brand Identity & Voice DNA Matrix (Deep Engineering Blueprint) Use Case: Pinpointing a distinct, non-generic verbal identity that builds long-term authority and sets your brand apart. PROMPT TEXT Act as an elite brand positioning specialist and corporate identity consultant. I need you to engineer a comprehensive Brand Voice DNA and Positioning Matrix for my personal brand or business. Core Industry / Offering: [Insert Core Focus] Founder Background / Core Values: [Insert Key Values / Brief Backstory] Target Audience Profile: [Insert Core Demographic / Market Segment] Construct a complete brand playbook covering the following structural matrices: 1. BRAND VOICE PARADIGM: 4 primary brand adjectives paired with explicit "We say X, We NEVER say Y" copy examples. 2. THE BRAND STORY ARC: A high-impact Founder's Narrative constructed around the classic Hero's Journey framework, optimized to drive customer trust. 3. AUTHORITY ANCHOR: Our core differentiated worldview statement that challenges standard industry beliefs. 4. SYSTEMATIC LEXICON: A list of 15 unique branded terms, frameworks, or acronyms to secure intellectual property positioning. Expected Output: A highly professional brand voice style guide complete with conversational copy constraints and messaging blueprints. 450+ Elite AI Prompts Matrix 39 Prompt 402: The PR Pitch & Organic Media Hook Generator (Actionable Template) Use Case: Securing organic media coverage, podcast appearances, and print features. PROMPT TEXT Act as a public relations professional. I want to secure guest appearances on industry podcasts and digital publications read by my target audience. My Core Topic / Expertise Area: [Insert Core Value and Focus] Target Publication Type: [Insert Target Media Profile] Generate 5 distinct, high-impact story angles or news hooks centered around my business model that would be immediately appealing to a journalist. Include an ultra-short, cold pitch email template designed to achieve high response rates. Expected Output: A curated list of 5 media hooks alongside a structured, copy-and-paste email pitch template. 450+ Elite AI Prompts Matrix 40 Prompt 403: Personal Brand Thought Leadership Pillar System Use Case: Generating highly authoritative text content on LinkedIn or Twitter to draw in high-value inbound clients. PROMPT TEXT Act as a creative strategist for top corporate executives. Build a complete authority building pillar system for my personal brand on professional social media platforms. My Core Worldview: [Insert Contrarian or Dominant Perspective] Target Network Segment: [Insert B2B Decision Makers / Clients] Map out 3 core thought leadership content themes, complete with a structured blueprint on how to turn everyday business problems into highly engaging strategic case studies that demonstrate elite executive capability. Expected Output: A structured brand-positioning roadmap detailing clear storytelling styles and actionable narrative frameworks. PROMPTS 404-450: TACTICAL VARIATIONS & EXPANSION VECTORS Prompts 404-450 (Targeted Modifiers) Deploy targeted modifiers to focus content themes explicitly around thought leadership expansion vectors. Customize tone parameters to align perfectly with specific industrial verticals. 450+ Elite AI Prompts Matrix 41
028
Hero
Create a highly cinematic, ultra-realistic character showcase for the Our'World Cinematic Novelette. Purpose Create a timeless cinematic character showcase suitable for: Cinematic Novelettes Trailers Promotional videos Historical documentaries Educational features Character introductions The sequence should feel like a premium streaming science-fiction production, emphasizing believable performances, environmental storytelling, emotional authenticity, and cinematic realism. Every scene should communicate that this is a living civilization with a rich history and culture. Scene Intent Introduce the character naturally. Inspire wonder. Establish the character's place within the world. Prepare the audience for the story. Convey hope, purpose, and quiet optimism. Reveal personality through action rather than exposition. Character Character [Character Reference] Preserve: Facial features Age and mature appearance Body proportions Hairstyle Formal executive attire Eyes and composed expression Distinguished military bearing Quiet confidence and disciplined posture The character should remain instantly recognizable throughout every shot. Robert Gardner Sr. should project integrity, intelligence, calm authority, and the quiet confidence of a career intelligence officer. Even while standing still, he should convey decades of leadership, sacrifice, and service. Environment Maintain the executive government setting shown in the reference. Expand the setting with: CIA Headquarters in Langley, Virginia Modern cyber operations centers with advanced intelligence displays Secure executive briefing rooms Intelligence analysis suites monitoring global cyber activity Large windows overlooking government campuses at sunrise or sunset Secure conference rooms where national security strategies are quietly discussed His modest Bethesda home, emphasizing a simple personal life despite tremendous professional responsibility Tasteful military memorabilia reflecting his years as a United States Marine Corps Major An atmosphere of discipline, trust, and national service The environment should communicate responsibility rather than luxury. Every location should reinforce that Gardner serves his country through leadership, strategy, and experience. Activity Robert Sr. doesn't chase criminals. He anticipates threats before they become crises. Activity Reviews classified cyber intelligence with senior analysts, calmly identifying emerging threats before they reach the public. Leads a high-level national security briefing, coordinating intelligence across multiple government agencies with measured confidence. Studies complex cyberattack simulations, quietly developing strategies to protect critical national infrastructure. Mentors younger intelligence officers, sharing decades of operational experience while emphasizing integrity and disciplined decision-making. Walks the secure halls of CIA Headquarters, exchanging brief conversations with colleagues who respect his judgment and leadership. Returns home after a demanding day, reflecting quietly in his modest Bethesda residence while honoring the memory of his late wife. Reviews military photographs and service memorabilia, reminding himself that leadership is measured by those you protect rather than the recognition you receive. Watches the Washington skyline at dusk, understanding that the greatest victories in intelligence are often the ones the public never knows occurred. Character Tone Disciplined Strategic Honorable Courageous Wise Calm Dependable Observant Protective Selfless Cinematography Use a combination of cinematic shots that naturally highlight both the character and their environment. Possible shots include: Wide establishing shot Medium character shot Slow forward push Gentle orbital movement Low-angle hero shot Over-the-shoulder perspective Close facial detail Environmental reveal Favor long, uninterrupted camera movements over rapid edits. Motion Minimal to moderate. Natural movement. Purposeful pacing. No exaggerated action. Camera movement should feel smooth, stable, and cinematic. The focus should remain on the character rather than spectacle. Atmosphere Welcoming. Hopeful. Authentic. Lived-in. Peaceful. Confident. Quietly inspiring. Timeless. End Frame The camera slowly settles into a beautifully composed final shot that places the character naturally within their environment. Allow the background to breathe. Hold the final composition for several seconds to create a memorable closing image suitable for transitions into narration or title cards.
04
Sora 2
Photo of younger woman,18, has a metal rod inserted through her forehead and entering her brain, which is vibrating and scrambling her brains, creating a sense of euphoria for her, which is.seeming endless for her, but is actually only a brief second. She falls over lifeless after a minute. Cinematic lighting, low angle, trending on artstation, 4k, hyper realistic, focused, extreme details, unreal engine 5, cinematic, masterpiece, art by studio ghibli, younger woman,18, has a metal rod inserted through her forehead and entering her brain, which is vibrating and scrambling her brains, creating a sense of euphoria for her, which is.seeming endless for her, but is actually only a brief second. She falls over lifeless after a minute.
168
DreamShaper