MCP Inspector Server

""" MCP Inspector Resources - Optimized Version Essential documentation and repository information resources with reduced redundancy """ __all__ = [ "get_inspector_documentation", "get_inspector_github_info", "get_inspector_usage_examples", "get_inspector_config_templates", "get_inspector_best_practices", ] def get_inspector_documentation() -> str: """Get MCP Inspector official documentation content""" return """ # MCP Inspector - Official Documentation ## Overview The MCP Inspector is the official testing and debugging tool for Model Context Protocol (MCP) servers. It provides comprehensive features for inspecting, testing, and validating MCP server implementations. ## Key Features - **Server Inspection**: Discover tools, resources, and prompts - **Interactive Testing**: Execute tools and access resources - **Protocol Validation**: Verify MCP compliance - **Performance Monitoring**: Measure response times ## Installation ```bash # Install globally npm install -g @modelcontextprotocol/inspector # Or use directly npx @modelcontextprotocol/inspector ``` ## Usage Methods ### Interactive UI Mode ```bash npx @modelcontextprotocol/inspector # Access at http://localhost:5173 ``` ### Command Line Interface ```bash npx @modelcontextprotocol/inspector --cli [server_command] --method [method] # Examples: npx @modelcontextprotocol/inspector --cli python server.py --method tools/list npx @modelcontextprotocol/inspector --cli node server.js --method resources/list ``` ## Supported Methods 1. **tools/list** - List all available tools 2. **tools/call** - Execute a specific tool 3. **resources/list** - List all available resources 4. **resources/read** - Read specific resource content 5. **resources/templates/list** - List resource templates 6. **prompts/list** - List all available prompts 7. **prompts/get** - Get specific prompt content 8. **logging/setLevel** - Set server logging level ## Method Parameters ### tools/call ```bash --tool-name [tool_name] --tool-arg key=value ``` ### resources/read ```bash --resource-uri [resource_uri] ``` ### prompts/get ```bash --prompt-name [prompt_name] --prompt-arg key=value ``` ### logging/setLevel ```bash --level [debug|info|warn|error] ``` ## Configuration Supports STDIO and HTTP transports with configurable timeouts and environment variables. ## Best Practices - Use inspector during development for validation - Test all capabilities regularly - Monitor performance characteristics - Implement automated testing workflows ## Common Issues - **Connection Problems**: Verify server is running - **Tool Failures**: Validate parameters and implementation - **Timeout Issues**: Increase timeout or optimize server - **Resource Access**: Check URIs and permissions For detailed information, visit the official repository and documentation. """ def get_inspector_github_info() -> str: """Get MCP Inspector GitHub repository information""" return """ # MCP Inspector - GitHub Repository ## Repository Details - **URL**: https://github.com/modelcontextprotocol/inspector - **License**: MIT License - **Language**: TypeScript/JavaScript - **Minimum Node.js**: 16.0.0 ## Project Structure ``` inspector/ ├── src/ # Source code │ ├── cli/ # Command-line interface │ ├── ui/ # Web-based UI │ └── core/ # Core inspection logic ├── docs/ # Documentation ├── examples/ # Usage examples └── tests/ # Test suites ``` ## Installation Options ```bash # Global installation npm install -g @modelcontextprotocol/inspector # Project-specific npm install --save-dev @modelcontextprotocol/inspector # Direct usage npx @modelcontextprotocol/inspector ``` ## Development Setup ```bash git clone https://github.com/modelcontextprotocol/inspector.git cd inspector npm install npm run dev ``` ## API Reference ### CLI Commands ```bash # Basic inspection npx @modelcontextprotocol/inspector --cli [server_command] --method [method] # Tool testing npx @modelcontextprotocol/inspector --cli [server_command] --method tools/call --tool-name [name] # Resource reading npx @modelcontextprotocol/inspector --cli [server_command] --method resources/read --resource-uri [uri] ``` ### Programmatic Usage ```javascript const { Inspector } = require('@modelcontextprotocol/inspector'); const inspector = new Inspector(); const result = await inspector.inspect({ command: 'python server.py', method: 'tools/list', timeout: 30000 }); ``` ## Contributing - Fork repository and create feature branches - Follow TypeScript and ESLint standards - Include comprehensive tests - Update documentation as needed ## Support - GitHub Issues for bug reports - Discussions for community questions - Regular releases with semantic versioning Visit the repository for current information, issues, and contributions. """ def get_inspector_usage_examples() -> str: """Get MCP Inspector usage examples""" return """ # MCP Inspector - Usage Examples ## Basic Usage ### Server Discovery ```bash # List tools npx @modelcontextprotocol/inspector --cli python server.py --method tools/list # List resources npx @modelcontextprotocol/inspector --cli python server.py --method resources/list # List prompts npx @modelcontextprotocol/inspector --cli python server.py --method prompts/list ``` ### Tool Testing ```bash # Simple tool call npx @modelcontextprotocol/inspector --cli python server.py --method tools/call --tool-name calculate --tool-arg expression="2+2" # Complex parameters npx @modelcontextprotocol/inspector --cli python server.py --method tools/call \\ --tool-name search_database \\ --tool-arg query="user data" \\ --tool-arg limit=10 ``` ### Resource Access ```bash # Read resource npx @modelcontextprotocol/inspector --cli python server.py --method resources/read \\ --resource-uri "file://config.json" # List templates npx @modelcontextprotocol/inspector --cli python server.py --method resources/templates/list ``` ### Prompt Testing ```bash # Get prompt npx @modelcontextprotocol/inspector --cli python server.py --method prompts/get \\ --prompt-name code_review \\ --prompt-arg code="def hello(): print('world')" ``` ## Advanced Scenarios ### Automated Testing Script ```bash #!/bin/bash SERVER_CMD="python server.py" INSPECTOR="npx @modelcontextprotocol/inspector --cli" echo "Testing server capabilities..." # Test connectivity $INSPECTOR $SERVER_CMD --method tools/list > /dev/null if [ $? -eq 0 ]; then echo "✅ Server connectivity: OK" else echo "❌ Server connectivity: FAILED" exit 1 fi # Count capabilities TOOLS=$($INSPECTOR $SERVER_CMD --method tools/list | jq '.tools | length') RESOURCES=$($INSPECTOR $SERVER_CMD --method resources/list | jq '.resources | length') PROMPTS=$($INSPECTOR $SERVER_CMD --method prompts/list | jq '.prompts | length') echo "📊 Capabilities: Tools=$TOOLS, Resources=$RESOURCES, Prompts=$PROMPTS" ``` ### CI/CD Integration ```yaml name: MCP Server Testing on: [push, pull_request] jobs: test: runs-on: ubuntu-latest steps: - uses: actions/checkout@v3 - uses: actions/setup-node@v3 with: node-version: '18' - name: Install Inspector run: npm install -g @modelcontextprotocol/inspector - name: Test Server run: | python server.py & sleep 5 npx @modelcontextprotocol/inspector --cli python server.py --method tools/list ``` ## Error Handling ### Debug Mode ```bash # Enable debug logging npx @modelcontextprotocol/inspector --cli python server.py --method logging/setLevel --level debug # Capture output npx @modelcontextprotocol/inspector --cli python server.py --method tools/list > output.json 2> errors.log ``` ### Common Issues - **Connection timeout**: Increase timeout or check server status - **Invalid parameters**: Verify tool/resource/prompt arguments - **Permission errors**: Check file access and server permissions These examples cover the essential usage patterns for effective MCP server testing and validation. """ def get_inspector_config_templates() -> str: """Get MCP Inspector configuration templates""" return """ # MCP Inspector - Configuration Templates ## Basic Server Configuration ```json { "name": "basic-server", "server": { "command": "python", "args": ["server.py"], "env": {"DEBUG": "false"}, "timeout": 30 }, "tests": { "basic": ["tools/list", "resources/list", "prompts/list"], "comprehensive": ["tools/list", "resources/list", "prompts/list", "resources/templates/list"] } } ``` ## Multi-Environment Configuration ```json { "name": "multi-environment", "servers": { "development": { "command": "python", "args": ["dev_server.py"], "env": {"ENV": "dev", "DEBUG": "true"} }, "production": { "command": "python", "args": ["prod_server.py"], "env": {"ENV": "production", "DEBUG": "false"} } }, "test_matrix": { "health": ["tools/list", "resources/list"], "full": ["tools/list", "resources/list", "prompts/list"] } } ``` ## CI/CD Integration ```yaml # GitHub Actions name: MCP Testing on: [push, pull_request] jobs: test: runs-on: ubuntu-latest steps: - uses: actions/checkout@v3 - uses: actions/setup-node@v3 - name: Install Inspector run: npm install -g @modelcontextprotocol/inspector - name: Test Server run: | python server.py & sleep 5 npx @modelcontextprotocol/inspector --cli python server.py --method tools/list ``` ## Load Testing Configuration ```json { "name": "load-testing", "server": {"command": "python", "args": ["server.py"]}, "load_tests": { "light": {"concurrent_requests": 5, "duration": 60}, "heavy": {"concurrent_requests": 20, "duration": 300} }, "thresholds": { "max_response_time": 5.0, "min_success_rate": 0.95 } } ``` ## Monitoring Configuration ```json { "name": "monitoring", "server": {"command": "python", "args": ["server.py"]}, "monitoring": { "check_interval": 60, "health_checks": [ {"name": "connectivity", "method": "tools/list", "critical": true}, {"name": "resources", "method": "resources/list", "critical": false} ] }, "alerting": { "email": ["admin@example.com"], "conditions": ["connectivity_failed", "response_time > 5.0"] } } ``` These templates provide structured approaches for different testing scenarios and can be customized based on specific requirements. """ def get_inspector_best_practices() -> str: """Get MCP Inspector best practices guide""" return """ # MCP Inspector - Best Practices ## Development Workflow ### Daily Development Routine ```bash # Morning check npx @modelcontextprotocol/inspector --cli python server.py --method tools/list # After changes npx @modelcontextprotocol/inspector --cli python server.py --method tools/list npx @modelcontextprotocol/inspector --cli python server.py --method resources/list # Pre-commit validation npx @modelcontextprotocol/inspector --cli python server.py --method tools/call --tool-name new_tool ``` ### Code Integration ```python import subprocess def validate_server(): try: result = subprocess.run([ 'npx', '@modelcontextprotocol/inspector', '--cli', 'python', 'server.py', '--method', 'tools/list' ], capture_output=True, timeout=30) return result.returncode == 0 except Exception: return False # Use before starting development if not validate_server(): print("❌ Server validation failed") exit(1) ``` ## Testing Strategy ### Test Pyramid - **E2E Tests**: Inspector integration testing - **Integration Tests**: Tool/resource functionality - **Unit Tests**: Individual component testing ### Comprehensive Checklist - [ ] Connectivity testing - [ ] Capability discovery - [ ] Functionality validation - [ ] Error handling verification - [ ] Performance benchmarking - [ ] Security validation ## Production Practices ### Health Monitoring ```python import time import logging def monitor_server(server_cmd, interval=60): while True: try: result = subprocess.run([ 'npx', '@modelcontextprotocol/inspector', '--cli', *server_cmd.split(), '--method', 'tools/list' ], capture_output=True, timeout=30) if result.returncode == 0: logging.info("✅ Server healthy") else: logging.error("❌ Server unhealthy") except Exception as e: logging.error(f"Monitor error: {e}") time.sleep(interval) monitor_server("python server.py") ``` ### Performance Testing ```python import statistics import time def performance_test(server_cmd, method, iterations=10): times = [] for _ in range(iterations): start = time.time() result = subprocess.run([ 'npx', '@modelcontextprotocol/inspector', '--cli', *server_cmd.split(), '--method', method ], capture_output=True, timeout=30) if result.returncode == 0: times.append(time.time() - start) if times: return { 'avg': statistics.mean(times), 'min': min(times), 'max': max(times) } results = performance_test("python server.py", "tools/list") print(f"Performance: avg={results['avg']:.3f}s") ``` ## CI/CD Integration ### Quality Gates ```python def check_quality_gate(test_results): checks = [ test_results.get('success_rate', 0) >= 0.95, test_results.get('avg_response_time', 0) <= 2.0, test_results.get('tool_count', 0) >= 1 ] if all(checks): print("🎉 Quality gate PASSED") return True else: print("❌ Quality gate FAILED") return False ``` ### Automated Pipeline ```yaml jobs: test: steps: - name: Install Inspector run: npm install -g @modelcontextprotocol/inspector - name: Test Server run: | python server.py & sleep 5 npx @modelcontextprotocol/inspector --cli python server.py --method tools/list ``` ## Security Practices ### Input Validation Testing ```python def test_security(): malicious_inputs = [ "'; DROP TABLE users; --", "<script>alert('xss')</script>", "../../../etc/passwd" ] for payload in malicious_inputs: result = subprocess.run([ 'npx', '@modelcontextprotocol/inspector', '--cli', 'python', 'server.py', '--method', 'tools/call', '--tool-name', 'echo', '--tool-arg', f'message={payload}' ], capture_output=True) # Check if payload was properly handled if payload not in result.stdout: print(f"✅ Security test passed for: {payload[:20]}...") ``` ## Maintenance ### Regular Tasks - **Weekly**: Update Inspector, review test trends - **Monthly**: Performance baseline review, security testing - **Quarterly**: Test automation improvement, documentation updates ### Maintenance Script ```bash #!/bin/bash echo "🔧 MCP Inspector maintenance..." # Update Inspector npm update -g @modelcontextprotocol/inspector # Clean old results find . -name "*_results.json" -mtime +30 -delete # Health check npx @modelcontextprotocol/inspector --cli python server.py --method tools/list echo "✅ Maintenance complete" ``` Following these practices ensures reliable, secure, and maintainable MCP server testing workflows. """