Skip to main content

CodePrism MCP Server Overview

CodePrism provides a production-ready MCP server that enables AI assistants to understand and navigate codebases through the Model Context Protocol. This page covers the MCP-specific implementation, tools, and integration details.

New to CodePrism? See the Introduction for a general overview, or check the Architecture for technical design details.

MCP Integration Features​

🔌 Complete MCP Compliance​

  • Resources: Repository files, code symbols, and graph data
  • Tools: 20 production-ready code analysis and navigation tools
  • Prompts: Code analysis templates and workflow guidance
  • Real-time Updates: Live resource and tool updates

🚀 Production-Ready Capabilities​

  • 100% Tool Success Rate - All 20 tools tested and production-ready
  • Semantic Parameter Support - Human-readable names instead of cryptic IDs
  • Environment Variable Integration - Automatic repository detection
  • Context-Enhanced Responses - Source code snippets included

MCP Server Configuration​

Server Capabilities​

{
  "name": "codeprism-mcp-server",
  "version": "0.3.4",
  "capabilities": {
    "resources": {
      "subscribe": true,
      "listChanged": true
    },
    "tools": {
      "listChanged": true
    },
    "prompts": {
      "listChanged": false
    }
  }
}

Transport Protocol​

  • JSON-RPC 2.0 over stdio
  • Request/Response Model for tools and resource access
  • Structured Error Handling with detailed context
  • Real-time Notifications for resource changes

Available Resources​

Repository Resources​

  • codeprism://repository/ - Repository root information and metadata
  • codeprism://repository/stats - Comprehensive repository statistics
  • codeprism://repository/config - Configuration and settings
  • codeprism://repository/tree - Complete file tree structure
  • codeprism://repository/file/{path} - Individual file content with analysis

Graph & Symbol Resources​

  • codeprism://graph/repository - Graph structure and statistics
  • codeprism://symbols/functions - All function symbols in the repository
  • codeprism://symbols/classes - All class symbols in the repository
  • codeprism://symbols/variables - All variable symbols in the repository
  • codeprism://symbols/modules - All module symbols in the repository

Analysis Resources​

  • codeprism://metrics/quality_dashboard - Code quality metrics and technical debt assessment
  • codeprism://architecture/layers - Architectural layer structure identification
  • codeprism://architecture/patterns - Detected design patterns and structures
  • codeprism://architecture/dependencies - High-level dependency analysis

Production Tools (20 Available)​

Core Navigation & Understanding (4 tools)​

  1. repository_stats - Get comprehensive statistics about the repository
  2. explain_symbol - Provide detailed explanation of a code symbol with context
  3. trace_path - Find execution paths between two code symbols
  4. find_dependencies - Analyze dependencies for a code symbol or file

Search & Discovery (4 tools)​

  1. search_symbols - Search for symbols by name pattern (with regex support)
  2. search_content - Search across all content including documentation and comments
  3. find_files - Find files by name or path pattern
  4. content_stats - Get statistics about indexed content

Advanced Analysis (11 tools)​

  1. find_unused_code - Graph-based unused code detection with confidence scoring
  2. analyze_security - Advanced vulnerability detection with CVSS scoring
  3. analyze_performance - Performance analysis with time complexity detection
  4. analyze_api_surface - API surface analysis with versioning compliance
  5. analyze_complexity - Calculate complexity metrics (cyclomatic, maintainability index)
  6. trace_data_flow - Bidirectional data flow tracing with transformation tracking
  7. analyze_transitive_dependencies - Complete dependency chain analysis with cycle detection
  8. detect_patterns - Identify design patterns and architectural structures
  9. trace_inheritance - Python inheritance hierarchy analysis with MRO support
  10. analyze_decorators - Comprehensive Python decorator analysis with framework detection
  11. find_duplicates - Duplicate code pattern detection with similarity scoring

Workflow & Orchestration (4 tools)​

  1. suggest_analysis_workflow - Get intelligent analysis recommendations for specific goals
  2. batch_analysis - Execute multiple analysis tools in parallel with result aggregation
  3. optimize_workflow - Optimize analysis workflows based on usage patterns
  4. find_references - Find all references to a symbol across the codebase

Available Prompts​

Repository Overview Prompt​

Name: repository_overview
Description: Generate a comprehensive overview of the repository structure and contents
Arguments:

  • focus_area (optional): Area to focus on (architecture, dependencies, entry_points, etc.)

Code Analysis Prompt​

Name: code_analysis
Description: Analyze code quality, patterns, and potential improvements
Arguments:

  • file_pattern (optional): File pattern to focus analysis on
  • analysis_type (optional): Type of analysis (quality, security, performance, architecture)

Debug Assistance Prompt​

Name: debug_assistance
Description: Help debug issues in the codebase with contextual information
Arguments:

  • issue_description (required): Description of the issue or error
  • affected_files (optional): Files related to the issue

Client Integration​

Claude Desktop​

Add to your Claude Desktop configuration (claude_desktop_config.json):

{
  "mcpServers": {
    "codeprism": {
      "command": "codeprism",
      "args": ["--mcp"],
      "env": {
        "CODEPRISM_PROJECT_ROOT": "/path/to/your/project"
      }
    }
  }
}

VS Code​

{
  "mcp.servers": [
    {
      "name": "codeprism",
      "command": "codeprism",
      "args": ["--mcp"],
      "workspaceFolder": "${workspaceFolder}"
    }
  ]
}

Cursor​

{
  "mcp.servers": {
    "codeprism": {
      "command": "codeprism",
      "args": ["--mcp", "--project-root", "${workspaceFolder}"]
    }
  }
}

Usage Patterns​

Basic Tool Usage​

// Example tool call via MCP
{
  "method": "tools/call",
  "params": {
    "name": "repository_stats",
    "arguments": {
      "include_metrics": true
    }
  }
}

Resource Access​

// Example resource access
{
  "method": "resources/read",
  "params": {
    "uri": "codeprism://repository/stats"
  }
}

Workflow Optimization​

Use the suggest_analysis_workflow tool to get intelligent recommendations for your analysis goals, then use batch_analysis for efficient parallel execution.

Next Steps​

Ready to use CodePrism? Follow the Installation Guide to get started with MCP integration.