IL2CPP Dump Analyzer Agentic RAG MCP System

A cutting-edge Agentic Retrieval-Augmented Generation (RAG) system for analyzing IL2CPP dump.cs files from Unity games. This system implements the Model Context Protocol (MCP) server specification with intelligent agentic capabilities that provide automated task orchestration, context-aware analysis, and enhanced workflow execution within the MCP framework.

🚀 Key Features

🤖 Agentic Intelligence Layer

• Intelligent MCP Tool Orchestration: Automatically selects and chains MCP tools for complex analysis tasks
• Context-Aware Processing: Preserves analysis context and results across multiple MCP tool calls
• Smart Task Decomposition: Breaks complex IL2CPP analysis requests into manageable subtasks
• Result Synthesis: Intelligently aggregates and correlates results from multiple MCP tools
• Adaptive Caching: Implements smart caching strategies that learn from usage patterns
• Performance Optimization: Monitors and optimizes agentic workflows for speed and efficiency

🔍 Advanced IL2CPP Analysis

• Semantic Code Processing: Specialized IL2CPPCodeChunker preserves code context and meaning
• Vector-Powered Search: Uses Xenova's Transformers.js with all-MiniLM-L6-v2 model (384-dimensional embeddings)
• Supabase Vector Database: High-performance vector search with pgvector extension
• Hash-based Change Detection: Intelligent change detection to avoid reprocessing unchanged files
• Metadata Extraction: Comprehensive assembly metadata, version tracking, and compilation analysis

🛠️ Comprehensive Tool Suite (21 MCP Tools)

• Search Tools (3): Semantic code search, metadata search, and advanced querying
• Analysis Tools (15): MonoBehaviour discovery, class hierarchies, design patterns, dependencies, and asset analysis
• Generation Tools (3): C# wrapper generation, method stubs, and Unity MonoBehaviour templates

🔌 MCP Integration Excellence

• Official MCP SDK: Full compliance with Model Context Protocol specification v1.12.0
• Multiple Transports: Stdio, HTTP, and Server-Sent Events (SSE) support
• Robust Validation: Zod schemas ensure comprehensive parameter validation
• Advanced Error Handling: Intelligent error recovery with detailed logging and retry mechanisms
• Resource Management: Efficient memory usage and connection pooling

Prerequisites

• Node.js 18.x or higher
• TypeScript (for development)
• Supabase account (required for vector database storage)
• IL2CPP dump.cs file from a Unity game

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/il2cpp-dump-analyzer-mcp.git
   cd il2cpp-dump-analyzer-mcp
   

2. Install dependencies:

   npm install
   

3. Set up Supabase database:

   • Create a new Supabase project at supabase.com
   • Run the SQL commands in supabase-setup.sql in the Supabase SQL editor
   • This creates the required tables with pgvector extension for vector storage

4. Configure environment variables:

   cp simple.env .env
   

   Update the .env file with your configuration:

   # Core Configuration
   NODE_ENV=production
   DUMP_FILE_PATH=./dump.cs
   EMBEDDING_MODEL=Xenova/all-MiniLM-L6-v2
   LOG_LEVEL=info
   
   # Supabase Configuration (Required)
   SUPABASE_URL=your_supabase_project_url
   SUPABASE_KEY=your_supabase_anon_key
   SUPABASE_TABLE_NAME=il2cpp_documents
   
   # MCP Server Configuration
   MCP_SERVER_PORT=3000
   MCP_SERVER_HOST=0.0.0.0
   

5. Build the project:

   npm run build
   

Usage

Environment Variables

All available environment variables including agentic configuration:

# Core Configuration
NODE_ENV=production|development|test
DUMP_FILE_PATH=./dump.cs
EMBEDDING_MODEL=Xenova/all-MiniLM-L6-v2
LOG_LEVEL=error|warn|info|debug

# Supabase Configuration
SUPABASE_URL=your_supabase_project_url
SUPABASE_KEY=your_supabase_anon_or_service_key
SUPABASE_TABLE_NAME=il2cpp_documents

# MCP Server Configuration
MCP_SERVER_PORT=3000
MCP_SERVER_HOST=0.0.0.0

# Agentic Configuration
AGENTIC_MODE=true                    # Enable intelligent tool orchestration
CONTEXT_CACHE_SIZE=1000             # Maximum cached contexts
TOOL_CHAIN_MAX_DEPTH=5              # Maximum tool chaining depth
INTELLIGENT_CACHING=true            # Enable smart result caching
CONTEXT_PERSISTENCE=true            # Enable context persistence across calls
PERFORMANCE_OPTIMIZATION=true      # Enable performance learning and optimization
ADAPTIVE_BATCHING=true              # Enable adaptive request batching

Quick Start

1. Place your IL2CPP dump.cs file in the root directory (or specify path in .env)

2. Start the MCP server:

   npm start
   

3. The server will automatically:

   • Parse the IL2CPP dump.cs file
   • Extract classes, methods, enums, and interfaces
   • Generate semantic embeddings using Xenova Transformers.js
   • Store vectors in Supabase with hash-based change detection
   • Start the MCP server with stdio transport

4. Connect with MCP clients:

   • Claude Desktop: Add to MCP configuration
   • Other MCP clients: Use stdio transport connection

MCP Client Configuration

Claude Desktop Configuration

Add to your Claude Desktop MCP configuration file:

{
  "mcpServers": {
    "il2cpp-analyzer": {
      "command": "node",
      "args": ["./bin/il2cpp-mcp-stdio.js"],
      "cwd": "/path/to/il2cpp-dump-analyzer-mcp"
    }
  }
}

Alternative: Direct Node.js Execution

# Run the MCP server directly
node ./bin/il2cpp-mcp-stdio.js

# Or use npm script
npm run mcp:stdio

🛠️ MCP Tools and Resources

The server provides 21 comprehensive MCP tools organized into three categories for IL2CPP analysis and code generation:

🔍 Search Tools (3 tools)

1. search_code - Semantic Code Search

Advanced semantic search through IL2CPP code with intelligent filtering.

Parameters:

• query (string, required): The search query
• filter_type (string, optional): Filter by entity type (class, method, enum, interface)
• filter_namespace (string, optional): Filter by namespace
• filter_monobehaviour (boolean, optional): Filter to only MonoBehaviour classes
• top_k (number, optional, default: 5): Number of results to return

Example:

// Find all Player-related classes with semantic search
search_code({ query: "Player movement controller", filter_type: "class", top_k: 10 })

2. search_metadata - Metadata Search

Search through extracted metadata with flexible filtering and advanced options.

Parameters:

• query (string, required): Search query
• search_type (string, optional): Type of search (assembly, type, method, field)
• assembly_name (string, optional): Filter by assembly name
• unity_version (string, optional): Filter by Unity version
• use_regex (boolean, optional): Enable regex pattern matching
• max_results (number, optional): Maximum results to return

3. query_metadata - Advanced Metadata Querying

Complex metadata queries with aggregations and cross-references.

Parameters:

• filters (object, optional): Complex filtering criteria
• aggregations (array, optional): Aggregation operations (count, group_by, statistics)
• cross_reference (object, optional): Cross-reference analysis configuration

🔬 Analysis Tools (15 tools)

4. find_monobehaviours - Unity Component Discovery

Find and analyze MonoBehaviour classes for Unity component analysis.

Parameters:

• query (string, optional): Optional search query to filter MonoBehaviours
• top_k (number, optional, default: 10): Number of results to return

5. find_class_hierarchy - Class Inheritance Analysis

Analyze class inheritance relationships and structure with detailed metadata.

Parameters:

• class_name (string, required): The name of the class to analyze
• include_methods (boolean, optional, default: true): Include methods in the output

6. find_enum_values - Enum Definition Extraction

Extract enum definitions and their values with type information.

Parameters:

• enum_name (string, required): The name of the enum to find values for

7. analyze_dependencies - Dependency Mapping

Comprehensive dependency analysis with circular dependency detection.

Parameters:

• class_name (string, required): Target class to analyze dependencies for
• analysis_type (enum, optional): Type of analysis (incoming, outgoing, bidirectional, circular)
• depth (number, optional, default: 3): How deep to traverse dependency chains (1-5)
• include_system_types (boolean, optional): Include Unity/System dependencies

8. find_cross_references - Cross-Reference Analysis

Find all references to specific code entities across the codebase.

Parameters:

• target_name (string, required): Name of the target entity
• target_type (enum, required): Type of entity (class, method, field, property, event)
• reference_type (enum, optional): Type of references (usage, inheritance, implementation)
• max_results (number, optional): Maximum number of references to return

9. find_design_patterns - Design Pattern Detection

Detect common design patterns with confidence scoring and examples.

Parameters:

• pattern_types (array, required): Array of patterns to detect (singleton, observer, factory, strategy, etc.)
• confidence_threshold (number, optional): Minimum confidence level (0.1-1.0)
• include_partial_matches (boolean, optional): Include partial pattern implementations

10. extract_metadata - Assembly Metadata Extraction

Extract comprehensive assembly metadata, version information, and compilation flags.

Parameters:

• content (string, optional): IL2CPP dump content
• file_path (string, optional): Path to IL2CPP dump file
• include_generic_instantiations (boolean, optional): Include generic type instantiations
• validate_structure (boolean, optional): Validate extracted metadata structure

11. analyze_type_hierarchies - Type System Analysis

Analyze complex type hierarchies including generics and constraints.

Parameters:

• type_name (string, required): Target type to analyze
• include_constraints (boolean, optional): Include generic constraints
• max_depth (number, optional): Maximum hierarchy depth to analyze

12. analyze_generic_types - Generic Type Analysis

Specialized analysis of generic types and their instantiations.

Parameters:

• type_name (string, required): Generic type to analyze
• include_instantiations (boolean, optional): Include all instantiations
• constraint_analysis (boolean, optional): Analyze type constraints

13. analyze_type_dependencies - Type Dependency Analysis

Analyze dependencies between types with detailed relationship mapping.

Parameters:

• type_name (string, required): Target type for dependency analysis
• dependency_type (enum, optional): Type of dependencies to analyze
• include_indirect (boolean, optional): Include indirect dependencies

14. analyze_type_compatibility - Type Compatibility Analysis

Analyze type compatibility and conversion possibilities.

Parameters:

• source_type (string, required): Source type name
• target_type (string, required): Target type name
• include_implicit_conversions (boolean, optional): Include implicit conversions

15. track_assembly_metadata - Assembly Version Tracking

Track assembly metadata changes and version comparisons.

Parameters:

• tracking_id (string, required): Unique tracking identifier
• comparison_mode (enum, optional): Comparison mode (full, incremental, diff)
• include_dependencies (boolean, optional): Include dependency tracking

16. analyze_asset_references - Asset Reference Analysis

Analyze asset references and dependencies within IL2CPP code.

Parameters:

• asset_type (string, optional): Type of assets to analyze
• include_missing_references (boolean, optional): Include missing asset references
• reference_depth (number, optional): Depth of reference analysis

17. find_unused_assets - Unused Asset Detection

Find potentially unused assets based on IL2CPP code analysis.

Parameters:

• asset_types (array, optional): Types of assets to check for usage
• exclude_patterns (array, optional): Patterns to exclude from analysis
• confidence_threshold (number, optional): Confidence threshold for unused detection

18. analyze_asset_dependencies - Asset Dependency Analysis

Comprehensive analysis of asset dependencies and circular references.

Parameters:

• include_circular_dependencies (boolean, optional): Include circular dependency detection
• asset_type_filter (array, optional): Filter by specific asset types
• dependency_depth (number, optional): Maximum dependency depth to analyze

⚙️ Generation Tools (3 tools)

19. generate_class_wrapper - C# Class Wrapper Generation

Generate C# wrapper classes from IL2CPP class definitions with full type fidelity.

Parameters:

• class_name (string, required): Name of the IL2CPP class to generate wrapper for
• include_methods (boolean, optional): Include method implementations
• include_properties (boolean, optional): Include property implementations
• generate_interfaces (boolean, optional): Generate interface definitions
• custom_namespace (string, optional): Custom namespace for generated code
• unity_version (string, optional): Target Unity version

Example:

// Generate comprehensive wrapper for Player class
generate_class_wrapper({
  class_name: "Player",
  include_methods: true,
  generate_interfaces: true,
  custom_namespace: "Game.Wrappers"
})

20. generate_method_stubs - Method Stub Generation

Generate method stubs and interfaces from IL2CPP class definitions.

Parameters:

• class_name (string, required): Name of the IL2CPP class to generate method stubs for
• method_filter (string, optional): Regex pattern to match specific methods
• include_documentation (boolean, optional): Include XML documentation comments
• include_error_handling (boolean, optional): Include error handling and validation
• generate_async (boolean, optional): Generate async/await patterns where applicable

21. generate_monobehaviour_template - Unity MonoBehaviour Template Generation

Generate Unity-ready MonoBehaviour scripts with common patterns and lifecycle methods.

Parameters:

• class_name (string, required): Name of the IL2CPP MonoBehaviour class
• template_type (string, optional): Template type (basic, ui, gameplay, system)
• include_lifecycle_methods (boolean, optional): Include Unity lifecycle methods
• include_unity_events (boolean, optional): Include UnityEvent implementations
• include_serialized_fields (boolean, optional): Include SerializeField attributes

🤖 Agentic Workflows and Usage Examples

Intelligent Tool Orchestration

The agentic layer automatically chains tools for complex analysis tasks:

// Example: Comprehensive Player class analysis
// The system automatically orchestrates multiple tools:

1. search_code({ query: "Player", filter_type: "class" })
2. find_class_hierarchy({ class_name: "Player" })
3. analyze_dependencies({ class_name: "Player", analysis_type: "bidirectional" })
4. find_design_patterns({ pattern_types: ["singleton", "observer"] })
5. generate_class_wrapper({ class_name: "Player", include_methods: true })

// All executed intelligently with context preservation

Context-Aware Analysis

// Example: MonoBehaviour ecosystem analysis
// The agent maintains context across tool calls:

1. find_monobehaviours({ query: "Enemy" })
   // Context: Found EnemyController, EnemyAI, EnemyHealth

2. analyze_dependencies({ class_name: "EnemyController" })
   // Context: Uses previous results to analyze dependencies

3. find_cross_references({ target_name: "EnemyController", target_type: "class" })
   // Context: Correlates with dependency analysis

4. generate_monobehaviour_template({ class_name: "EnemyController" })
   // Context: Uses all previous analysis for optimal template generation

Smart Caching and Performance

The agentic system implements intelligent caching:

• Result Caching: Frequently accessed analysis results are cached
• Context Persistence: Analysis context is preserved across sessions
• Performance Learning: The system learns from usage patterns to optimize future requests
• Adaptive Batching: Related tool calls are automatically batched for efficiency

Resources

The server exposes resources through the MCP resource system:

• il2cpp://{query}: Retrieves code snippets matching the query with intelligent ranking
  • Query parameters:
    • top_k: Number of results to return (default: 5)
    • filter_type: Filter by entity type (class, method, enum, interface)
    • filter_namespace: Filter by namespace
    • filter_monobehaviour: Filter to only include MonoBehaviour classes
    • context_aware: Enable context-aware result ranking (default: true)

Development

Development Mode

1. Run in development mode:

   npm run dev
   

2. Run tests:

   npm test
   npm run test:watch    # Watch mode
   npm run test:coverage # With coverage
   

3. Lint and format code:

   npm run lint
   npm run format
   

4. Build for production:

   npm run build
   

Testing

The project includes comprehensive Jest testing infrastructure with agentic component coverage:

• Unit Tests: All MCP tools, agentic components, and core functionality
• Integration Tests: Vector store operations, Supabase integration, and tool orchestration
• Performance Tests: Large file processing, embedding generation, and agentic workflow optimization
• Agentic Tests: Context management, tool selection, and intelligent orchestration
• Error Handling Tests: Edge cases, error scenarios, and recovery mechanisms

Run specific test suites:

npm run test:unit        # Unit tests only
npm run test:integration # Integration tests only
npm run test:performance # Performance tests only
npm run test:mcp-tools   # MCP tool-specific tests
npm run test             # Full test suite including agentic components

Agentic Component Testing

# Test agentic orchestration
npm run test -- --testPathPattern=agent

# Test intelligent tool selection
npm run test -- --testPathPattern=mcp-tool-selector

# Test context management
npm run test -- --testPathPattern=mcp-context-manager

Project Structure

src/
├── __tests__/              # Test files and test utilities
│   ├── setup.ts           # Jest test setup
│   ├── test-data.ts       # Mock IL2CPP data for testing
│   ├── agent/             # Agentic component tests
│   ├── integration/       # Integration tests
│   ├── mcp/               # MCP tool tests
│   └── *.test.ts          # Individual test files
├── agent/                 # Agentic Intelligence Layer
│   ├── mcp-orchestrator.ts # Intelligent tool orchestration
│   ├── mcp-context-manager.ts # Context management and persistence
│   ├── mcp-performance-optimizer.ts # Performance optimization
│   ├── mcp-response-synthesizer.ts # Multi-tool result synthesis
│   ├── mcp-tool-selector.ts # AI-driven tool selection
│   ├── types.ts           # Agentic type definitions
│   └── index.ts           # Agent exports
├── config/                 # Configuration utilities
├── database/              # Database connection and management
│   ├── connection-manager.ts # Database connection pooling
│   ├── enhanced-vector-store.ts # Enhanced vector operations
│   └── performance-monitor.ts # Database performance monitoring
├── embeddings/            # Embedding generation and vector storage
│   ├── chunker.ts         # IL2CPP-specific code chunking
│   ├── xenova-embeddings.ts # Xenova Transformers.js integration
│   ├── supabase-vector-store.ts # Supabase vector store implementation
│   └── vector-store.ts    # Main vector store interface
├── generator/             # Code generation infrastructure
│   ├── types.ts           # TypeScript interfaces for code generation
│   ├── base-generator.ts  # Abstract base class for generators
│   ├── template-engine.ts # Template engine integration
│   ├── class-wrapper-generator.ts # C# class wrapper generator
│   ├── method-stub-generator.ts # Method stub generator
│   ├── monobehaviour-generator.ts # Unity MonoBehaviour template generator
│   └── index.ts           # Generator exports
├── indexer/               # File indexing and processing
│   └── indexer.ts         # Main indexing logic with hash management
├── mcp/                   # MCP server implementation
│   ├── mcp-sdk-server.ts  # Main MCP server with all tools
│   ├── stdio-server.ts    # Stdio transport server
│   ├── tools/             # MCP tool implementations (21 tools)
│   │   ├── search-code-tool.ts # Semantic code search
│   │   ├── find-monobehaviours-tool.ts # MonoBehaviour discovery
│   │   ├── analyze-dependencies-tool.ts # Dependency analysis
│   │   ├── generate-class-wrapper-tool.ts # Code generation
│   │   ├── tool-registry.ts # Tool registration and metadata
│   │   └── ... (18 more tools)
│   └── types.ts           # MCP type definitions
├── metadata/              # Metadata analysis and extraction
│   └── type-analyzer.ts   # Advanced type analysis
├── monitoring/            # System monitoring and health
│   ├── health-service.ts  # Health monitoring
│   ├── metrics-service.ts # Performance metrics
│   └── lifecycle-manager.ts # Component lifecycle management
├── parser/                # IL2CPP dump file parsing
│   ├── il2cpp-parser.ts   # Main parser implementation
│   ├── enhanced-il2cpp-parser.ts # Enhanced parser with metadata
│   ├── advanced-parser.ts # Advanced parsing capabilities
│   └── index.ts           # Parser exports
├── performance/           # Performance optimization
│   ├── chunked-processor.ts # Chunked processing for large files
│   ├── batch-vector-store.ts # Batch vector operations
│   └── streaming-parser.ts # Streaming parser for memory efficiency
├── transport/             # Transport layer implementations
│   ├── http-transport.ts  # HTTP transport
│   ├── transport-factory.ts # Transport factory
│   └── index.ts           # Transport exports
└── utils/                 # Utility functions
    ├── hash-manager.ts    # File hash management
    ├── supabase-hash-manager.ts # Supabase-based hash storage
    ├── mcp-response-formatter.ts # Response formatting
    └── parameter-validator.ts # Parameter validation

bin/
└── il2cpp-mcp-stdio.js    # Executable MCP server binary

examples/                   # Code generation examples and documentation
├── README.md              # Examples overview
├── class-wrapper-example.md # Class wrapper generation examples
├── method-stubs-example.md # Method stub generation examples
└── monobehaviour-template-example.md # MonoBehaviour template examples

supabase-setup.sql          # Supabase database schema

🏗️ Architecture

Foundation Layer

1. Enhanced IL2CPP Parser: Advanced parsing with metadata extraction and type analysis
2. Semantic Code Chunker: Context-preserving chunking with IL2CPP-specific optimizations
3. Xenova Embeddings Engine: 384-dimensional embeddings using Transformers.js all-MiniLM-L6-v2
4. Supabase Vector Store: High-performance vector search with pgvector extension
5. Hash-based Change Detection: Intelligent file change tracking and incremental processing

Agentic Intelligence Layer

6. MCP Tool Orchestrator: Intelligent tool selection and workflow automation
7. Context Manager: Persistent context across tool calls with smart compression
8. Performance Optimizer: Real-time performance monitoring and optimization
9. Response Synthesizer: Multi-tool result aggregation and correlation
10. Tool Selector: AI-driven tool selection based on task requirements

MCP Server Layer

11. MCP SDK Server: Official MCP TypeScript SDK implementation with 21 specialized tools
12. Transport Layer: Multi-transport support (stdio, HTTP, SSE) with connection pooling
13. Validation Engine: Comprehensive Zod schema validation with error recovery
14. Resource Manager: Efficient memory management and connection handling

Data Flow with Agentic Enhancement

1. Input Processing: IL2CPP dump.cs file with intelligent preprocessing
2. Agentic Analysis: Smart task decomposition and tool selection
3. Parallel Processing: Concurrent parsing, chunking, and embedding generation
4. Vector Storage: Optimized storage in Supabase with intelligent indexing
5. Context-Aware Retrieval: Smart caching and context-aware search
6. Multi-Tool Orchestration: Automated tool chaining for complex analysis
7. Result Synthesis: Intelligent aggregation and correlation of results
8. Adaptive Learning: Performance optimization based on usage patterns

MCP SDK Integration

This project uses the official Model Context Protocol TypeScript SDK (@modelcontextprotocol/sdk) for full MCP compliance:

Key Features

• Standardized Protocol: Full MCP specification compliance
• Resource Templates: Expose IL2CPP data through MCP resources
• Tool Definitions: Comprehensive parameter validation using Zod schemas
• Stdio Transport: Optimized for desktop applications and command-line tools
• Error Handling: Robust error management with detailed logging
• Session Management: Stateful interactions with MCP clients

Transport Configuration

The server uses stdio transport only for optimal compatibility with:

• Claude Desktop
• Command-line MCP clients
• Desktop applications
• Development tools

Performance Considerations

• Incremental Processing: Hash-based change detection avoids reprocessing
• Efficient Chunking: Semantic-aware chunking preserves code meaning
• Vector Optimization: 384-dimensional embeddings balance quality and performance
• Database Indexing: Optimized Supabase queries with proper indexing
• Memory Management: Streaming processing for large dump files

Troubleshooting

Common Issues

1. Supabase Connection Errors

   • Verify SUPABASE_URL and SUPABASE_KEY in .env
   • Ensure pgvector extension is enabled
   • Check network connectivity

2. Embedding Generation Slow

   • First run downloads the model (~90MB)
   • Subsequent runs use cached model
   • Consider using faster hardware for large files

3. MCP Client Connection Issues

   • Verify stdio transport configuration
   • Check file permissions on bin/il2cpp-mcp-stdio.js
   • Ensure Node.js is in PATH

4. Memory Issues with Large Files

   • Increase Node.js memory limit: node --max-old-space-size=4096
   • Consider chunking very large dump files

5. Code Generation Issues

   • Class Not Found: Ensure the class exists in the IL2CPP dump and is properly indexed
   • Invalid Generated Code: Check Unity version compatibility and namespace conflicts
   • Missing Dependencies: Verify all required using statements are included
   • Type Resolution Errors: Ensure IL2CPP dump contains complete type information

6. MonoBehaviour Generation Issues

   • Not a MonoBehaviour: Verify the target class inherits from MonoBehaviour
   • Missing Unity Methods: Check Unity version compatibility for lifecycle methods
   • Serialization Issues: Ensure fields are properly marked as serializable

🐳 Docker Support

The IL2CPP Dump Analyzer MCP system includes comprehensive Docker support for easy deployment and development.

Quick Start with Docker

1. Setup Environment:

   # Linux/macOS
   ./docker-setup.sh
   
   # Windows PowerShell
   .\docker-setup.ps1
   

2. Start Production Environment:

   docker-compose --env-file .env.docker up -d
   

3. Start Development Environment:

   docker-compose -f docker-compose.dev.yml --env-file .env.docker.dev up -d
   

Docker Architecture

The system uses a multi-container architecture:

• IL2CPP MCP Server: Main application container with Xenova embeddings
• Supabase Database: PostgreSQL with pgvector extension
• Supabase REST API: PostgREST API gateway
• Kong Gateway: API gateway and routing (production)
• Supabase Studio: Database management UI (development)

Recent Docker Improvements

✅ Fixed Xenova Model Loading: Proper path resolution and timeout handling ✅ Enhanced Memory Management: Increased limits for model loading (4GB) ✅ Improved Startup Times: Extended health check periods (5 minutes) ✅ Better Error Handling: Retry logic and graceful failure recovery ✅ Volume Optimization: Named volumes for better cross-platform compatibility

Troubleshooting

If you encounter Docker issues, see DOCKER-TROUBLESHOOTING.md for detailed solutions.

🤝 Contributing

We welcome contributions to the IL2CPP Dump Analyzer Agentic RAG MCP System! Please follow these guidelines:

Development Guidelines

1. Fork the repository
2. Create a feature branch: git checkout -b feature/amazing-agentic-feature
3. Follow Test-Driven Development (TFD): Write tests before implementing features
4. Add comprehensive tests: Include unit, integration, and agentic component tests
5. Run the full test suite: npm test
6. Test agentic components: npm run test -- --testPathPattern=agent
7. Commit your changes: git commit -m 'Add amazing agentic feature'
8. Push to the branch: git push origin feature/amazing-agentic-feature
9. Open a Pull Request

Agentic Development Guidelines

When contributing to agentic components:

• Context Preservation: Ensure context is properly managed across tool calls
• Performance Optimization: Consider performance implications of intelligent workflows
• Tool Orchestration: Design workflows that intelligently chain MCP tools
• Error Recovery: Implement robust error handling and recovery mechanisms
• Adaptive Learning: Consider how the system can learn from usage patterns

Code Quality Standards

• TypeScript Strict Mode: Maintain full TypeScript compliance
• JSDoc Documentation: Document all functions with comprehensive JSDoc comments
• Zod Validation: Use Zod schemas for all input validation
• MCP Compliance: Ensure all MCP tools follow the official specification
• Agentic Patterns: Follow established patterns for agentic component development

License

MIT License - see LICENSE file for details.

🙏 Acknowledgments

• Model Context Protocol for the MCP specification and agentic framework foundation
• Xenova/Transformers.js for client-side embeddings and AI capabilities
• Supabase for vector database infrastructure and real-time capabilities
• TypeScript for type safety and development experience
• Unity Technologies for IL2CPP technology and game development innovation
• The open-source community for continuous inspiration and collaboration

Built with ❤️ for the Unity game development and reverse engineering community

Empowering developers with intelligent IL2CPP analysis through agentic AI and the Model Context Protocol