## Projects Organization - Create status-based folders: WIP/PAUSE/CONSTANT/CONCEPT/ARCHIVE - Move 17 projects to appropriate status folders - Delete obsolete README.md ### WIP (4 projects) - GroveEngine, SEO_Article_Generator, AISSIA, SecondVoice ### PAUSE (6 projects) - Warfactory, chinese_audio_tts_pipeline, MCP_Game_Asset_Pipeline - ocr_pdf_service, Essay_Writing_Tingting, shipping_strategy/ ### CONSTANT (3 projects) - ClassGen (Analysis + 2.0), Database_Cours_Chinois, civjdr ### CONCEPT (5 projects) - pokrovsk_last_day, pokrovsk_drone_command (NEW full design doc) - social_network_manager, vps_tunnel_china, Claude_Workflow_Optimization ### ARCHIVE (3 items) - MCP_Creative_Amplification, Backlog_9-10_Octobre_2025, LeBonCoup/ ## Tracking Files Updated - Status_Projets.md: Complete rewrite with current state (Nov 2025) - planning/TODO_data.md: Updated with new structure and all projects by status - CLAUDE.md: Updated relation status, Projects section, daily check stats ## Daily Check System - Add card ACTION-008: essay_writing_tingting - Update card_database.md: 21 total cards (15 Tingting, 3 Personal, 1 Family, 1 Tech, 1 Comm) 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
699 lines
16 KiB
Markdown
699 lines
16 KiB
Markdown
# MCP Game Asset Pipeline - Vision Masterclass
|
|
|
|
## TL;DR - Le Concept
|
|
|
|
**Pipeline créatif AI-powered pour génération et intégration automatique d'assets de jeu en temps réel, avec découverte émergente de gameplay.**
|
|
|
|
Un système où :
|
|
- Le jeu génère du contenu en réponse aux actions du joueur
|
|
- Les dessins vectoriels deviennent des assets pixel-perfect
|
|
- Claude Code orchestre DALL-E, Suno, et autres AIs via MCP
|
|
- Hot-reload permet l'intégration immédiate sans interruption
|
|
- Le gameplay émerge de la création
|
|
|
|
**Status:** Concept / Vision (pas encore implémenté)
|
|
**Application cible:** Warfactory (extensible à tout jeu modulaire)
|
|
|
|
---
|
|
|
|
## 🎯 Vision Globale
|
|
|
|
### Le Problème
|
|
Créer des assets de jeu (sprites, musique, SFX) est :
|
|
- Chronophage
|
|
- Requiert compétences techniques variées
|
|
- Pipeline lourd (design → création → intégration → test)
|
|
- Itération lente
|
|
- Limite la créativité par friction
|
|
|
|
### La Solution : MCP + AI + Hot-Reload
|
|
|
|
**Pipeline unifié orchestré par Claude Code via MCP :**
|
|
|
|
```
|
|
Intention créative (humain)
|
|
↓
|
|
Claude Code (analyse + orchestration)
|
|
↓ (MCP)
|
|
AIs spécialisées (DALL-E, Suno, etc.)
|
|
↓
|
|
Validation automatique
|
|
↓
|
|
Hot-reload dans le jeu
|
|
↓
|
|
Gameplay émergent
|
|
```
|
|
|
|
---
|
|
|
|
## 🏗️ Architecture
|
|
|
|
### Layer 1 : Game Engine (Warfactory/Godot)
|
|
```
|
|
- Système modulaire + JSON config
|
|
- Hot-reload de modules
|
|
- Event system (découvertes, combos, etc.)
|
|
- MCP client intégré
|
|
```
|
|
|
|
### Layer 2 : MCP Protocol
|
|
```
|
|
Serveurs MCP spécialisés :
|
|
├── dalle-mcp # Génération images
|
|
├── suno-mcp # Génération musique
|
|
├── music-analyzer # Analyse audio (Essentia)
|
|
├── asset-validator # Validation technique
|
|
└── discovery-engine # Émergence gameplay
|
|
```
|
|
|
|
### Layer 3 : Orchestration (Claude Code)
|
|
```
|
|
- Analyse contexte créatif
|
|
- Génération prompts optimisés
|
|
- Validation multi-critères
|
|
- Itération automatique
|
|
- Gestion workflow complet
|
|
```
|
|
|
|
### Layer 4 : Integration (n8n optionnel)
|
|
```
|
|
- Centralisation workflows
|
|
- Logs et monitoring
|
|
- Storage et versioning
|
|
- Notifications
|
|
- Gestion coûts API
|
|
```
|
|
|
|
---
|
|
|
|
## 🎨 Use Cases Masterclass
|
|
|
|
### 1. Discovery Loop - Émergence Gameplay
|
|
|
|
**Concept :** Le jeu génère du nouveau contenu basé sur les actions du joueur.
|
|
|
|
```
|
|
IN-GAME EVENT:
|
|
Player découvre combo ["fire", "metal", "steam"]
|
|
dans "industrial_zone"
|
|
↓
|
|
MCP EVENT:
|
|
{
|
|
"type": "combo_discovered",
|
|
"elements": ["fire", "metal", "steam"],
|
|
"context": "industrial_zone",
|
|
"player_style": "aggressive_builder"
|
|
}
|
|
↓
|
|
CLAUDE CODE REASONING:
|
|
"Combo fire+metal+steam + zone industrielle + joueur agressif
|
|
→ Je génère une 'War Forge'
|
|
→ Stats: +production +défense -ressources
|
|
→ Style: Dark industrial fantasy"
|
|
↓
|
|
ASSET GENERATION:
|
|
- Building sprite (DALL-E)
|
|
- SFX forge + steam (ElevenLabs)
|
|
- Particle config (auto)
|
|
- JSON config
|
|
↓
|
|
HOT-RELOAD:
|
|
3 secondes → Nouveau bâtiment disponible in-game!
|
|
↓
|
|
PLAYER EXPERIENCE:
|
|
"WTF je viens de débloquer quelque chose d'inédit!"
|
|
```
|
|
|
|
**Impact :** Gameplay unique, rejouabilité infinie, émergence réelle.
|
|
|
|
---
|
|
|
|
### 2. Vectorial Workflow - Dessin → Asset
|
|
|
|
**Concept :** Transformer dessins vectoriels en assets pixel-perfect.
|
|
|
|
```
|
|
PHASE 1: CAPTURE
|
|
Tingting/Alexis dessine sur iPad (Procreate/Concepts)
|
|
- Dessin vectoriel propre
|
|
- Annotations textuelles ("glow here", "aggressive look", "faction red")
|
|
↓
|
|
PHASE 2: ANALYSE (MCP → Claude Code)
|
|
Parse SVG:
|
|
- Formes: "3 cylindres, 2 sphères, base rectangulaire"
|
|
- Style: "Lignes épaisses = bold, courbes = organique"
|
|
- Annotations: "Glow effect, style agressif, couleur faction rouge"
|
|
- Proportions: Ratios exacts
|
|
↓
|
|
PHASE 3: PROMPT GENERATION
|
|
"Game asset, isometric view, [formes détectées],
|
|
[style inféré], [annotations],
|
|
low-poly aesthetic, clean edges, game-ready,
|
|
style guide: Warfactory industrial"
|
|
↓
|
|
PHASE 4: GENERATION
|
|
DALL-E génère l'asset
|
|
↓
|
|
PHASE 5: VALIDATION
|
|
- Overlay vectoriel sur résultat DALL-E
|
|
- Compare proportions (tolérance 5%)
|
|
- Vérifie annotations respectées
|
|
- Score de match: 0-100%
|
|
↓
|
|
PHASE 6: ITERATION
|
|
Si score < 85%:
|
|
- Corrections précises dans prompt
|
|
- Régénération
|
|
- Re-validation
|
|
Loop jusqu'à satisfaction
|
|
↓
|
|
PHASE 7: FINALIZATION
|
|
Asset parfait + metadata + versioning
|
|
→ Ready for game
|
|
```
|
|
|
|
**Impact :** Vision artistique humaine + Qualité technique IA = Hybrid creativity sans friction.
|
|
|
|
---
|
|
|
|
### 3. Pipeline Complet : Concept → Playable
|
|
|
|
**Commande unique :**
|
|
```
|
|
"Claude, ajoute une unité 'Steampunk Tank' à Warfactory"
|
|
```
|
|
|
|
**Exécution automatique :**
|
|
|
|
```
|
|
1. DESIGN DOC
|
|
- Analyse gameplay existant
|
|
- Propose stats équilibrées
|
|
- Définit rôle tactique
|
|
- Écrit description lore
|
|
|
|
2. ASSETS VISUELS
|
|
- Sprite idle
|
|
- Sprite move
|
|
- Sprite attack
|
|
- Sprite death
|
|
- Particle effects config
|
|
|
|
3. ASSETS AUDIO
|
|
- SFX engine
|
|
- SFX attack
|
|
- SFX destruction
|
|
- Musique thème (optionnel)
|
|
|
|
4. CODE
|
|
- Script unité Godot/Unity
|
|
- Behaviour tree IA
|
|
- Tests unitaires
|
|
|
|
5. INTEGRATION
|
|
- JSON config
|
|
- Balance pass
|
|
- Documentation
|
|
|
|
6. DELIVERY
|
|
- Commit git avec preview
|
|
- PR avec screenshots
|
|
- Changelog généré
|
|
|
|
TEMPS TOTAL: ~5-10 minutes
|
|
(vs plusieurs jours manuellement)
|
|
```
|
|
|
|
**Impact :** De l'idée au prototype jouable en une commande.
|
|
|
|
---
|
|
|
|
### 4. Asset Variants Automatiques
|
|
|
|
**Un prompt → Multiples variations :**
|
|
|
|
```
|
|
INPUT:
|
|
"Génère un building 'Factory' pour Warfactory"
|
|
|
|
OUTPUT AUTOMATIQUE:
|
|
├── factory_base.png # Version de base
|
|
├── factory_damaged_1.png # Légèrement endommagé
|
|
├── factory_damaged_2.png # Très endommagé
|
|
├── factory_destroyed.png # Ruines
|
|
├── factory_upgraded_t2.png # Amélioration tier 2
|
|
├── factory_upgraded_t3.png # Amélioration tier 3
|
|
├── factory_faction_red.png # Variant faction rouge
|
|
├── factory_faction_blue.png # Variant faction bleue
|
|
└── factory_spritesheet.png # Spritesheet assemblée
|
|
|
|
+ JSON config avec tous les états
|
|
+ Metadata (prompts, versions, style guide match)
|
|
```
|
|
|
|
**Impact :** Un asset → Ecosystem complet en une passe.
|
|
|
|
---
|
|
|
|
### 5. Adaptive Music System
|
|
|
|
**Concept :** Musique qui s'adapte au gameplay en temps réel.
|
|
|
|
```
|
|
PHASE 1: ANALYSE GAMEPLAY
|
|
Claude Code analyse la scène Warfactory:
|
|
- Rythme: Combat intense / Construction calme / Exploration
|
|
- Tension: Score de danger
|
|
- Progression: Early/Mid/Late game
|
|
|
|
PHASE 2: GENERATION LAYERS
|
|
Génère musique avec layers séparés:
|
|
├── base_ambient.mp3 # Layer ambiance (toujours actif)
|
|
├── rhythm_construction.mp3 # Layer construction
|
|
├── tension_combat.mp3 # Layer combat
|
|
└── climax_boss.mp3 # Layer moments clés
|
|
|
|
PHASE 3: ANALYSE AUDIO
|
|
Essentia analyse chaque layer:
|
|
- BPM synchronisé
|
|
- Tonalité cohérente
|
|
- Points de transition
|
|
- Energy levels
|
|
|
|
PHASE 4: INTEGRATION
|
|
Système audio adaptatif:
|
|
- Transitions seamless entre layers
|
|
- Crossfade selon état jeu
|
|
- Dynamic mixing
|
|
|
|
PHASE 5: TEST
|
|
Simule scenarios:
|
|
- Construction → Combat : Transition smooth?
|
|
- Combat → Victoire : Climax works?
|
|
```
|
|
|
|
**Impact :** OST unique pour chaque partie, adapté au style du joueur.
|
|
|
|
---
|
|
|
|
### 6. AI Art Director Persistant
|
|
|
|
**Concept :** Style guide auto-maintenu qui assure cohérence visuelle.
|
|
|
|
```
|
|
STYLE GUIDE WARFACTORY:
|
|
{
|
|
"palette": {
|
|
"primary": ["#2C3E50", "#E74C3C", "#95A5A6"],
|
|
"factions": {
|
|
"red": ["#C0392B", "#E74C3C"],
|
|
"blue": ["#2980B9", "#3498DB"]
|
|
}
|
|
},
|
|
"style": {
|
|
"aesthetic": "low-poly industrial",
|
|
"line_weight": "bold",
|
|
"lighting": "top-down 45°",
|
|
"saturation": "medium-low",
|
|
"contrast": "high"
|
|
},
|
|
"proportions": {
|
|
"building_height": "1.2-1.8x base",
|
|
"unit_scale": "0.3-0.5x building",
|
|
"detail_level": "medium"
|
|
},
|
|
"rules": [
|
|
"Edges must be clean",
|
|
"No gradients",
|
|
"Max 3 main colors per asset",
|
|
"Consistent perspective"
|
|
]
|
|
}
|
|
```
|
|
|
|
**Workflow avec Art Director :**
|
|
|
|
```
|
|
1. Nouveau asset généré
|
|
↓
|
|
2. Claude Code vérifie conformité style guide
|
|
↓
|
|
3. Score de match: 0-100%
|
|
↓
|
|
4. Si < 80% : Rejection + corrections spécifiques
|
|
↓
|
|
5. Si 80-95% : Ajustements mineurs
|
|
↓
|
|
6. Si > 95% : Approved
|
|
↓
|
|
7. Style guide mis à jour si nouveau pattern validé manuellement
|
|
```
|
|
|
|
**Impact :** Cohérence visuelle automatique, pas de "ça ressemble pas au jeu".
|
|
|
|
---
|
|
|
|
### 7. Collaborative Creation (Alexis + Tingting)
|
|
|
|
**Workflow couple :**
|
|
|
|
```
|
|
SCENARIO 1: Tingting dessine concept
|
|
↓
|
|
Upload vers n8n
|
|
↓
|
|
Claude Code analyse + génère variations
|
|
↓
|
|
Notification Telegram: "3 variations prêtes"
|
|
↓
|
|
Tingting/Alexis votent sur préférée
|
|
↓
|
|
Production automatique assets finaux
|
|
↓
|
|
Commit git avec "Co-created by Tingting & Alexis"
|
|
|
|
SCENARIO 2: Brainstorming vocal
|
|
↓
|
|
Audio transcrit (Whisper)
|
|
↓
|
|
Claude extrait idées
|
|
↓
|
|
Génère moodboard visuel + audio
|
|
↓
|
|
Review ensemble
|
|
↓
|
|
Sélection collaborative
|
|
↓
|
|
Pipeline complet se lance
|
|
```
|
|
|
|
**Impact :** Création de couple, outils pour collaborer artistiquement.
|
|
|
|
---
|
|
|
|
## 🛠️ Stack Technique
|
|
|
|
### MCP Servers à créer
|
|
|
|
#### 1. dalle-mcp
|
|
```
|
|
Tools exposés:
|
|
- generate_image(prompt, style_guide, size)
|
|
- generate_variations(base_image, count)
|
|
- analyze_image(image) → description
|
|
- compare_images(img1, img2) → similarity score
|
|
```
|
|
|
|
#### 2. suno-mcp
|
|
```
|
|
Tools exposés:
|
|
- generate_music(prompt, duration, style)
|
|
- generate_layers(base_track, layer_types)
|
|
- extend_track(track, duration)
|
|
```
|
|
|
|
#### 3. music-analyzer-mcp
|
|
```
|
|
Tools exposés (Essentia):
|
|
- analyze_audio(file) → tempo, key, energy, mood
|
|
- check_sync(track1, track2) → compatibility
|
|
- find_transitions(track) → transition points
|
|
```
|
|
|
|
#### 4. asset-validator-mcp
|
|
```
|
|
Tools exposés:
|
|
- validate_sprite(image, specs) → pass/fail + issues
|
|
- check_style_guide(image, guide) → score
|
|
- generate_spritesheet(images, layout)
|
|
```
|
|
|
|
#### 5. discovery-engine-mcp
|
|
```
|
|
Tools exposés:
|
|
- analyze_combo(elements, context) → suggestions
|
|
- generate_concept(discovery_event) → design doc
|
|
- balance_stats(entity, game_state) → balanced stats
|
|
```
|
|
|
|
### Integration Points
|
|
|
|
#### Warfactory (Godot)
|
|
```gdscript
|
|
# MCP Event emitter
|
|
func on_discovery(combo: Array, context: String):
|
|
var event = {
|
|
"type": "combo_discovered",
|
|
"elements": combo,
|
|
"context": context,
|
|
"player_stats": get_player_stats()
|
|
}
|
|
MCPClient.emit_event(event)
|
|
|
|
# Hot-reload listener
|
|
func _on_mcp_asset_ready(asset_data):
|
|
load_json_config(asset_data.config_path)
|
|
reload_module(asset_data.module_name)
|
|
```
|
|
|
|
#### n8n Workflows
|
|
```
|
|
Workflow 1: Asset Generation Pipeline
|
|
├── Webhook trigger (MCP event)
|
|
├── Claude Code API call
|
|
├── DALL-E generation
|
|
├── Storage (S3/local)
|
|
├── Validation
|
|
└── Hot-reload trigger
|
|
|
|
Workflow 2: Vectorial to Sprite
|
|
├── Watch folder (SVG uploads)
|
|
├── SVG analysis
|
|
├── Prompt generation
|
|
├── DALL-E call
|
|
├── Validation loop
|
|
└── Asset finalization
|
|
|
|
Workflow 3: Discovery System
|
|
├── Game event webhook
|
|
├── Claude reasoning
|
|
├── Multi-service generation (parallel)
|
|
├── Assembly
|
|
└── Game integration
|
|
```
|
|
|
|
---
|
|
|
|
## 📊 Metrics & Optimization
|
|
|
|
### Tracking
|
|
|
|
```
|
|
Asset Generation Metrics:
|
|
├── Generation time (target: < 30s)
|
|
├── Iteration count (target: < 3)
|
|
├── Style guide match score (target: > 90%)
|
|
├── Manual rejection rate (target: < 10%)
|
|
└── API costs per asset
|
|
|
|
Gameplay Metrics:
|
|
├── Discovery events triggered
|
|
├── Player engagement with generated content
|
|
├── Asset usage in-game (used vs unused)
|
|
└── Player retention correlation
|
|
|
|
Economic Metrics:
|
|
├── Cost per asset (DALL-E, Suno, etc.)
|
|
├── Local vs Cloud usage ratio
|
|
├── Monthly budget tracking
|
|
└── ROI: Time saved vs cost
|
|
```
|
|
|
|
### Learning System
|
|
|
|
```
|
|
ML Model learns:
|
|
- Prompts qui fonctionnent best
|
|
- Patterns de rejection
|
|
- Préférences utilisateur
|
|
- Style émergent
|
|
|
|
Result:
|
|
- Prompts optimisés automatiquement
|
|
- "Alexis style" qui émerge
|
|
- Moins d'itérations nécessaires
|
|
- Meilleure qualité over time
|
|
```
|
|
|
|
---
|
|
|
|
## 🗺️ Roadmap d'Implémentation
|
|
|
|
### Phase 1 : Proof of Concept (1-2 semaines)
|
|
```
|
|
☐ Serveur MCP DALL-E basique
|
|
☐ Claude Code: "Génère 1 building Warfactory"
|
|
☐ Hot-reload manuel dans Godot
|
|
☐ Validation: "Ça marche end-to-end"
|
|
```
|
|
|
|
### Phase 2 : Pipeline Manuel (2-4 semaines)
|
|
```
|
|
☐ Vectorial analysis (SVG parsing)
|
|
☐ Style guide system
|
|
☐ Validation automatique
|
|
☐ Asset variants generation
|
|
☐ Import workflow optimisé
|
|
```
|
|
|
|
### Phase 3 : Automation (4-6 semaines)
|
|
```
|
|
☐ Watch folders (auto-detection)
|
|
☐ Hot-reload automatique
|
|
☐ n8n workflows complets
|
|
☐ Metrics & monitoring
|
|
☐ Cost management
|
|
```
|
|
|
|
### Phase 4 : Discovery Loop (6-8 semaines)
|
|
```
|
|
☐ Game event system
|
|
☐ Discovery engine MCP
|
|
☐ Real-time generation
|
|
☐ Balance system
|
|
☐ Full émergence gameplay
|
|
```
|
|
|
|
### Phase 5 : Masterclass Features (8-12 semaines)
|
|
```
|
|
☐ Adaptive music system
|
|
☐ AI Art Director persistant
|
|
☐ Collaborative tools (Tingting + Alexis)
|
|
☐ ML optimization
|
|
☐ Multi-game support
|
|
```
|
|
|
|
---
|
|
|
|
## 💡 Extensions Possibles
|
|
|
|
### Au-delà de Warfactory
|
|
|
|
**Applicabilité générale :**
|
|
- N'importe quel jeu avec système modulaire
|
|
- Game jams (prototyping ultra-rapide)
|
|
- Indie devs (réduction drastique coûts)
|
|
- Educational (apprendre game design)
|
|
|
|
### Autres domaines créatifs
|
|
|
|
**Film/Animation :**
|
|
- Storyboard → Concept art
|
|
- Scene generation
|
|
- Character variants
|
|
|
|
**UI/UX Design :**
|
|
- Wireframe → High-fidelity mockups
|
|
- Component variants
|
|
- Design system maintenance
|
|
|
|
**Architecture/Product Design :**
|
|
- Sketch → 3D render
|
|
- Material variants
|
|
- Context adaptation
|
|
|
|
---
|
|
|
|
## ⚠️ Challenges & Limitations
|
|
|
|
### Techniques
|
|
- Latence API (DALL-E ~10-30s par image)
|
|
- Coûts API (à surveiller)
|
|
- Qualité variable (itérations nécessaires)
|
|
- Style consistency (pas toujours parfait)
|
|
|
|
### Design
|
|
- Over-generation risk (trop de contenu = dilution)
|
|
- Balance gameplay (assets générés doivent rester équilibrés)
|
|
- Cohérence narrative (lore peut devenir incohérent)
|
|
|
|
### Solutions
|
|
- Cache intelligent (assets similaires)
|
|
- Budget caps avec alertes
|
|
- Human-in-the-loop validation
|
|
- Style guide strict + Art Director AI
|
|
- Narrative constraints dans prompts
|
|
|
|
---
|
|
|
|
## 🎓 Pourquoi c'est Masterclass
|
|
|
|
### Innovation technique
|
|
✓ MCP comme bridge entre AI et game engine
|
|
✓ Hot-reload pour itération temps réel
|
|
✓ Multi-AI orchestration
|
|
|
|
### Innovation gameplay
|
|
✓ Émergence réelle (pas scriptée)
|
|
✓ Contenu unique par partie
|
|
✓ Player-driven content generation
|
|
|
|
### Innovation workflow
|
|
✓ Hybrid human-AI creativity
|
|
✓ Vectorial → Pixel-perfect pipeline
|
|
✓ Collaborative tools (couple)
|
|
|
|
### Impact
|
|
✓ 10-100x réduction temps création
|
|
✓ Accessibilité (non-artists peuvent créer)
|
|
✓ Nouveaux types de jeux possibles
|
|
|
|
---
|
|
|
|
## 📚 Références & Inspirations
|
|
|
|
**Concepts similaires :**
|
|
- Procedural generation (No Man's Sky, Minecraft)
|
|
- AI Dungeon (émergence narrative)
|
|
- GitHub Copilot (AI-assisted creation)
|
|
|
|
**Technologies clés :**
|
|
- MCP (Model Context Protocol)
|
|
- DALL-E / Midjourney (image generation)
|
|
- Suno / Udio (music generation)
|
|
- Essentia (audio analysis)
|
|
- n8n (workflow automation)
|
|
|
|
**Philosophie :**
|
|
- Tools for thought (amplifier créativité humaine)
|
|
- Emergent gameplay (systèmes simples → comportements complexes)
|
|
- Democratization (rendre accessible à tous)
|
|
|
|
---
|
|
|
|
## 🚀 Next Steps
|
|
|
|
**Pour démarrer :**
|
|
|
|
1. **Valider l'intérêt** : Ce concept te fait vraiment vibrer ?
|
|
2. **Définir scope initial** : Warfactory only ou plus large ?
|
|
3. **Allouer temps** : Side project ou focus principal ?
|
|
4. **Budget API** : Combien tu veux investir en tests ?
|
|
|
|
**Si GO :**
|
|
- Phase 1 proof of concept
|
|
- 1-2 semaines d'expérimentation
|
|
- Si ça marche → Continue
|
|
- Si pas concluant → Pivot ou archive
|
|
|
|
---
|
|
|
|
**Document créé le :** 2025-10-14
|
|
**Auteur :** Alexis (avec Claude Code)
|
|
**Status :** Vision / Concept
|
|
**Prochaine review :** TBD
|
|
|
|
---
|
|
|
|
*"Le jeu ne se joue plus, il se crée en jouant."*
|