feat: transform to AI Development Cockpit with LangGraph orchestration

Major architectural transformation from dual-domain platform to AI agent orchestration system that democratizes coding.

## Vision Change
- FROM: Dual-domain LLM platform (SwaggyStacks/ScientiaCapital)
- TO: AI Development Cockpit that transforms noobs into software engineering managers
- Target users: Beginners who want to build software by managing AI agent teams

## Architecture
- LangGraph 1.0 multi-agent state machine orchestration
- Human-in-the-loop approval gates (architecture, testing, deployment)
- Service-to-service integration with ai-cost-optimizer microservice
- 90% cost savings via intelligent LLM routing

## New Structure
- src/orchestrator/ - LangGraph state machine and workflow
  - graph.ts - Main state graph with approval gates
  - AgentOrchestrator.ts - High-level orchestration API
- src/agents/ - AI agent implementations
  - BaseAgent.ts - Abstract base class with cost-optimized thinking
  - CodeArchitect.ts - System design and architecture planning
- src/services/cost-optimizer/ - Integration with ai-cost-optimizer service
  - CostOptimizerClient.ts - Service-to-service HTTP client
- src/types/orchestrator.ts - Complete type definitions for orchestration system

## Dependencies Added
- @langchain/core ^0.3.28
- @langchain/langgraph ^0.2.42
- @langchain/openai ^0.3.18
- langchain ^0.3.11
- uuid ^11.0.3

## Documentation Updates
- README.md: Complete rewrite with democratization vision
- CLAUDE.md: LangGraph architecture and agent system docs
- .env.example: Updated for ai-cost-optimizer microservice integration
- package.json: Renamed from dual-domain-llm-platform to ai-development-cockpit

## 5 Core Agents (Initial Implementation)
1. CodeArchitect - System designer (implemented)
2. BackendDeveloper - API & database builder (placeholder)
3. FrontendDeveloper - UI/UX builder (placeholder)
4. Tester - Quality assurance (placeholder)
5. DevOpsEngineer - Deployment & infrastructure (placeholder)

🚀 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

Author: ScientiaCapital

.env.example

@@ -49,15 +49,18 @@ NEXT_PUBLIC_SUPABASE_ANON_KEY="your_supabase_anon_key_here" # Required: Supabase
 SUPABASE_SERVICE_ROLE_KEY="your_service_role_key_here"    # Required: Admin operations & migrations
 
 # ----------------------------------------------------------------------------
-# COST OPTIMIZER CONFIGURATION (ai-cost-optimizer integration)
+# COST OPTIMIZER CONFIGURATION (ai-cost-optimizer microservice integration)
 # ----------------------------------------------------------------------------
-COST_OPTIMIZER_ENABLED="true"                             # Enable intelligent routing (true/false)
-COST_OPTIMIZER_DEFAULT_TIER="auto"                        # Routing strategy: auto | free | mid | premium
-COST_OPTIMIZER_COMPLEXITY_THRESHOLD="100"                 # Token count for simple vs complex (default: 100)
-COST_OPTIMIZER_DAILY_BUDGET="5.00"                        # Daily budget in USD (alerts when exceeded)
-COST_OPTIMIZER_MONTHLY_BUDGET="50.00"                     # Monthly budget in USD (hard limit)
-COST_OPTIMIZER_API_URL="http://localhost:3001"            # Internal API endpoint
-COST_OPTIMIZER_SAVINGS_TARGET="0.60"                      # Target 60% cost reduction
+# The ai-cost-optimizer is a SEPARATE microservice deployed independently
+# GitHub: https://github.com/ScientiaCapital/ai-cost-optimizer
+# This service routes AI requests to the most cost-effective provider
+# providing 90% cost savings (Gemini free tier → Claude Haiku → Premium)
+# ----------------------------------------------------------------------------
+COST_OPTIMIZER_API_URL="https://your-cost-optimizer-vercel-url.vercel.app"  # Required: ai-cost-optimizer service URL
+                                                           # Production: Your Vercel deployment URL
+                                                           # Development: http://localhost:3002 (if running locally)
+COST_OPTIMIZER_API_KEY="your_cost_optimizer_api_key_here" # Required: API key for cost optimizer service
+                                                           # Generate from ai-cost-optimizer dashboard
 
 # ----------------------------------------------------------------------------
 # RUNPOD CONFIGURATION (Serverless Chinese LLMs)
@@ -120,27 +123,36 @@ ENABLE_MARKETPLACE="true"                                 # Enable model marketp
 # ============================================================================
 # NOTES:
 # ============================================================================
-# 1. Cost Optimization Strategy:
-#    - 70% of queries route to Gemini Flash (FREE)
+# 1. Architecture Overview:
+#    - This is an AI DEVELOPMENT COCKPIT - orchestrates AI agents to build software
+#    - Integrates with ai-cost-optimizer microservice for 90% cost savings
+#    - LangGraph-based multi-agent system with human-in-the-loop approval gates
+#    - Transforms noobs into software engineering managers
+#
+# 2. Cost Optimization Strategy (via ai-cost-optimizer microservice):
+#    - 70% of queries route to Gemini Flash (FREE via ai-cost-optimizer)
 #    - 25% route to Claude Haiku (~$0.13/day)
-#    - 5% route to RunPod Chinese LLMs (~$0.05/day)
-#    - Expected monthly cost: $15-20 (vs $45-50 without optimization)
+#    - 5% route to Premium models when needed
+#    - Expected monthly cost: $4-5 (vs $45-50 without optimization)
+#
+# 3. Required API Keys (Minimum to start):
+#    - COST_OPTIMIZER_API_URL + API_KEY (ai-cost-optimizer service)
+#    - NEXT_PUBLIC_SUPABASE_URL + ANON_KEY (authentication & database)
 #
-# 2. Required API Keys (Minimum to start):
-#    - GOOGLE_API_KEY (free tier available)
-#    - ANTHROPIC_API_KEY (required for complex queries)
-#    - OPENROUTER_API_KEY (fallback routing)
-#    - NEXT_PUBLIC_SUPABASE_URL + ANON_KEY (authentication)
+# 4. ai-cost-optimizer Setup:
+#    - Deploy ai-cost-optimizer separately to Vercel
+#    - GitHub: https://github.com/ScientiaCapital/ai-cost-optimizer
+#    - Configure API keys in that service (GOOGLE_API_KEY, ANTHROPIC_API_KEY, etc.)
+#    - This cockpit only needs the cost optimizer's URL and API key
 #
-# 3. Optional but Recommended:
+# 5. Optional but Recommended:
 #    - RUNPOD_API_KEY (for Chinese LLM support)
 #    - REDIS_URL (for caching and performance)
-#    - COST_ALERT_WEBHOOK (for budget monitoring)
 #
-# 4. Security Best Practices:
+# 6. Security Best Practices:
 #    - Never commit this file with real keys
 #    - Use .env.local for actual credentials (gitignored)
 #    - Rotate keys regularly
-#    - Use service role keys only server-side
+#    - The ai-cost-optimizer handles LLM API keys - this service only talks to optimizer
 #
 # ============================================================================