Fix v0 server code extraction and add enhanced features

.claude/commands/todo_guide

@@ -0,0 +1,61 @@
+# Claude Code Todo System Guide
+
+## How Todos Actually Work
+
+The todo system in Claude Code stores todo lists as JSON files with UUID names in:
+```
+/home/gotime2022/.claude/todos/
+```
+
+Each session gets a unique UUID, for example:
+- `/home/gotime2022/.claude/todos/a64ca605-74cc-4915-afa3-19ac8c315b41.json`
+
+## The Problem
+
+1. Each new Claude Code session creates a NEW todo file
+2. Previous session todos are saved but not loaded
+3. This creates the illusion that todos don't persist
+
+## The Solution
+
+To access todos from previous sessions:
+
+1. List recent todo files:
+```bash
+ls -t /home/gotime2022/.claude/todos/*.json | head -10
+```
+
+2. Read a specific todo file:
+```
+Read(/home/gotime2022/.claude/todos/[UUID].json)
+```
+
+3. Find non-empty todo files:
+```bash
+for f in $(ls -t /home/gotime2022/.claude/todos/*.json | head -20); do 
+  size=$(wc -c < "$f")
+  if [ $size -gt 10 ]; then 
+    echo "File: $f (size: $size)"
+    cat "$f" | head -5
+  fi
+done
+```
+
+## Permissions Required
+
+Add these to `.claude/settings.json`:
+```json
+"Read(/home/gotime2022/.claude/todos/*.json)",
+"TodoRead()",
+"TodoWrite(*)"
+```
+
+## Current Session Todo Commands
+
+- `TodoRead()` - Read current session's todos
+- `TodoWrite(todos)` - Update current session's todos
+- `/todo` - Display todos in interactive mode
+
+## The Real Issue
+
+The todo system works but lacks session continuity. Each session is isolated, which breaks the workflow when you need to continue work from a previous session.

.env.example

@@ -1,46 +1,67 @@
-# DevLoopAI Environment Variables
+# MCP Server Environment Variables
 # Copy this file to .env and fill in your actual values
-# NEVER commit the .env file to version control
 
-# API Keys - REQUIRED
-ANTHROPIC_API_KEY=your_anthropic_api_key_here
-OPENAI_API_KEY=your_openai_api_key_here
-GEMINI_API_KEY=your_gemini_api_key_here
+# GitHub MCP Server
+GITHUB_TOKEN=your_github_personal_access_token
 
-# Supabase Configuration - REQUIRED
-SUPABASE_URL=your_supabase_project_url
-SUPABASE_ANON_KEY=your_supabase_anon_key
-SUPABASE_SERVICE_ROLE_KEY=your_supabase_service_role_key
+# OpenAI Tools MCP Server  
+OPENAI_API_KEY=your_openai_api_key
 
-# GitHub Configuration - OPTIONAL
-GITHUB_TOKEN=your_github_personal_access_token
-GITHUB_ORG=your_github_organization
+# Anthropic Comprehensive MCP Server
+ANTHROPIC_API_KEY=your_anthropic_api_key
 
-# Slack Configuration - OPTIONAL
-SLACK_BOT_TOKEN=your_slack_bot_token
-SLACK_APP_TOKEN=your_slack_app_token
+# Supabase MCP Server
+SUPABASE_URL=https://your-project.supabase.co
+SUPABASE_SERVICE_KEY=your_supabase_service_key
+SUPABASE_ACCESS_TOKEN=your_supabase_access_token
+
+# Vercel v0 MCP Server (for AI component generation)
+V0_API_KEY=your_v0_api_key
 
-# Redis Configuration - OPTIONAL
-REDIS_URL=redis://localhost:6379
+# Vercel Deploy MCP Server (for deployments)
+VERCEL_TOKEN=your_vercel_deploy_token
+
+# Gemini MCP Server
+GEMINI_API_KEY=your_gemini_api_key
+
+# Slack MCP Server
+SLACK_BOT_TOKEN=your_slack_bot_token
+SLACK_USER_TOKEN=your_slack_user_token
 
-# Application Configuration
-NODE_ENV=development
-NEXT_PUBLIC_API_URL=http://localhost:8080
-API_PORT=8080
-FRONTEND_PORT=3001
+# Redis MCP Server
+REDIS_HOST=localhost
+REDIS_PORT=6379
+REDIS_PASSWORD=optional_redis_password
 
-# Security
-JWT_SECRET=generate_a_strong_random_string_here
-SESSION_SECRET=generate_another_strong_random_string_here
+# Context7 MCP Server
+CONTEXT7_API_KEY=your_context7_api_key
 
-# Feature Flags
-ENABLE_MCP_SERVERS=true
-ENABLE_PARALLEL_EXECUTION=true
-ENABLE_AUTO_SAVE=true
+# Browserbase MCP Server
+BROWSERBASE_API_KEY=your_browserbase_api_key
+BROWSERBASE_PROJECT_ID=your_browserbase_project_id
 
-# Logging
-LOG_LEVEL=info
-LOG_FORMAT=json
+# Hostinger MCP Server
+HOSTINGER_API_KEY=your_hostinger_api_key
 
-# Database
-DATABASE_URL=postgresql://user:password@localhost:5432/devloop
+# MCP Server Ports (defaults shown)
+BRAVE_SEARCH_MCP_PORT=8003
+FILESYSTEM_MCP_PORT=8006
+MEMORY_MCP_PORT=8007
+VERCEL_V0_MCP_PORT=8010
+GITHUB_MCP_PORT=8011
+OPENAI_TOOLS_MCP_PORT=8012
+SUPABASE_MCP_PORT=8013
+GEMINI_MCP_PORT=8014
+ANTHROPIC_MCP_PORT=8015
+SEQUENTIAL_THINKING_MCP_PORT=8016
+SLACK_MCP_PORT=8017
+REDIS_MCP_PORT=8018
+CONTEXT7_MCP_PORT=8019
+DOCKER_MCP_PORT=8020
+EVERYTHING_MCP_PORT=8021
+FETCH_MCP_PORT=8022
+BROWSERBASE_MCP_PORT=8023
+HOSTINGER_MCP_PORT=8024
+VERCEL_DEPLOY_MCP_PORT=8025
+ROUTING_MCP_PORT=8026
+EOF < /dev/null

V0_SERVER_IMPROVEMENTS_SUMMARY.md

@@ -0,0 +1,84 @@
+# V0 Server Major Improvements Summary
+
+## Branch: feat/v0-server-major-improvements
+
+### Overview
+This branch contains significant improvements to the Vercel v0 MCP server to address code extraction issues and add new functionality for better file management and feature generation.
+
+## Key Improvements Made
+
+### 1. Fixed Code Extraction Issues
+- **Problem**: The v0 server was saving entire AI responses including markdown artifacts and explanatory text
+- **Solution**: Updated `_parse_generated_code` method with improved regex patterns to:
+  - Properly extract code from markdown code blocks
+  - Handle file markers like ` ```tsx file="path/to/file.ext" ` `
+  - Remove language identifiers and closing backticks
+  - Support multiple files in a single response
+
+### 2. Enhanced Multi-File Support
+- **Problem**: Generated files were all placed in the same directory regardless of specified paths
+- **Solution**: Updated `create_component_files` method to:
+  - Use the `path` field from file_info dict when provided
+  - Create proper directory structures based on file paths
+  - Handle paths that start with `/` or `./` correctly
+
+### 3. Added write_to_file Parameter
+- **Enhancement**: Added `write_to_file` parameter to `generate_component` tool
+- **Benefits**:
+  - Allows direct file creation during component generation
+  - Reduces the need for separate file creation steps
+  - Includes `target_directory` parameter for flexible file placement
+
+### 4. Implemented v0_update_file Function
+- **New Tool**: `v0_update_file` - Updates existing files in-place using v0
+- **Features**:
+  - Reads existing file content
+  - Improves it based on provided requirements
+  - Writes the improved code back to the file
+  - Maintains file structure and API compatibility
+
+### 5. Enhanced improve_component Function
+- **Improvements**:
+  - Added `write_to_file` parameter to directly update files
+  - Added `file_path` parameter for specifying which file to update
+  - Better error handling for file write operations
+
+### 6. Added v0_create_feature Function
+- **New Tool**: `v0_create_feature` - Creates complete features with multiple components
+- **Capabilities**:
+  - Generates multiple related components for a feature
+  - Creates optional API endpoints
+  - Ensures components work together with consistent interfaces
+  - Handles file creation for all generated components
+
+## Code Quality Improvements
+
+### Error Handling
+- Added try/catch blocks around file operations
+- Proper error messages when file creation fails
+- Partial success reporting when some files are created
+
+### Logging
+- Added detailed logging for debugging code extraction
+- Log file creation operations with sizes
+- Log raw v0 responses for troubleshooting
+
+### File Management
+- Consistent UTF-8 encoding with Unix line endings
+- Proper path handling for cross-platform compatibility
+- Directory creation with parents=True for nested structures
+
+## Testing Recommendations
+
+1. Test multi-file generation with v0
+2. Verify file paths are created correctly
+3. Test the v0_update_file function on existing components
+4. Test v0_create_feature with multiple components
+5. Verify markdown artifacts are properly removed
+
+## Next Steps
+
+1. Copy all fixes to devloop3 repository
+2. Test with real-world component generation scenarios
+3. Add more sophisticated error recovery mechanisms
+4. Consider adding component preview functionality