mcp-adr-analysis-server

#!/bin/bash # MCP Functionality Testing Script # Tests MCP server functionality using @modelcontextprotocol/inspector set -euo pipefail # Colors for output RED='\033[0;31m' GREEN='\033[0;32m' YELLOW='\033[1;33m' BLUE='\033[0;34m' NC='\033[0m' # No Color # Configuration SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" PROJECT_ROOT="$(dirname "$SCRIPT_DIR")" MCP_SERVER_PATH="$PROJECT_ROOT/dist/src/index.js" TEST_TIMEOUT=30 MCP_INSPECTOR_TIMEOUT=10 echo -e "${BLUE}🔍 MCP Functionality Testing${NC}" echo "===============================" # Function to print colored output print_status() { local status=$1 local message=$2 case $status in "INFO") echo -e "${BLUE}ℹ️ $message${NC}" ;; "SUCCESS") echo -e "${GREEN}✅ $message${NC}" ;; "WARNING") echo -e "${YELLOW}⚠️ $message${NC}" ;; "ERROR") echo -e "${RED}❌ $message${NC}" ;; esac } # Function to check prerequisites check_prerequisites() { print_status "INFO" "Checking prerequisites..." # Check if built server exists if [ ! -f "$MCP_SERVER_PATH" ]; then print_status "ERROR" "MCP server not found at $MCP_SERVER_PATH" print_status "INFO" "Building the project..." npm run build if [ ! -f "$MCP_SERVER_PATH" ]; then print_status "ERROR" "Failed to build MCP server" exit 1 fi fi # Check if server is executable if [ ! -x "$MCP_SERVER_PATH" ]; then print_status "WARNING" "Making MCP server executable..." chmod +x "$MCP_SERVER_PATH" fi # Check Node.js version NODE_VERSION=$(node --version) print_status "INFO" "Node.js version: $NODE_VERSION" print_status "SUCCESS" "Prerequisites check passed" } # Function to test basic MCP server health test_server_health() { print_status "INFO" "Testing MCP server health..." # Use gtimeout on macOS, timeout on Linux local timeout_cmd="timeout" if command -v gtimeout >/dev/null 2>&1; then timeout_cmd="gtimeout" elif ! command -v timeout >/dev/null 2>&1; then # Fallback: no timeout command print_status "WARNING" "No timeout command available, running without timeout..." if node "$MCP_SERVER_PATH" --test; then print_status "SUCCESS" "MCP server health check passed" return 0 else print_status "ERROR" "MCP server health check failed" return 1 fi return fi if $timeout_cmd $TEST_TIMEOUT node "$MCP_SERVER_PATH" --test; then print_status "SUCCESS" "MCP server health check passed" return 0 else print_status "ERROR" "MCP server health check failed" return 1 fi } # Function to test MCP tools listing using inspector CLI test_mcp_tools_list() { print_status "INFO" "Testing MCP tools listing..." # Use MCP Inspector CLI to list tools local tools_output local timeout_cmd="timeout" if command -v gtimeout >/dev/null 2>&1; then timeout_cmd="gtimeout" elif ! command -v timeout >/dev/null 2>&1; then timeout_cmd="" fi if [ -n "$timeout_cmd" ]; then tools_output=$($timeout_cmd $MCP_INSPECTOR_TIMEOUT npx @modelcontextprotocol/inspector --cli node "$MCP_SERVER_PATH" --method tools/list 2>/dev/null) else tools_output=$(npx @modelcontextprotocol/inspector --cli node "$MCP_SERVER_PATH" --method tools/list 2>/dev/null) fi if [ $? -eq 0 ] && [ -n "$tools_output" ]; then print_status "SUCCESS" "MCP tools list retrieved successfully" # Check for expected tools (based on actual MCP server implementation) local expected_tools=( "analyze_project_ecosystem" "get_architectural_context" "generate_adrs_from_prd" "generate_adr_todo" "compare_adr_progress" "analyze_content_security" "generate_content_masking" "suggest_adrs" "review_existing_adrs" "tool_chain_orchestrator" ) local found_tools=0 for tool in "${expected_tools[@]}"; do # Use jq to parse JSON if available, otherwise use grep on the "name" field if command -v jq >/dev/null 2>&1; then if echo "$tools_output" | jq -e ".tools[]? | select(.name == \"$tool\")" >/dev/null 2>&1; then print_status "SUCCESS" "Found tool: $tool" ((found_tools++)) else print_status "WARNING" "Tool not found: $tool" fi else # Fallback: look for the tool name in the JSON structure if echo "$tools_output" | grep -q "\"name\": \"$tool\""; then print_status "SUCCESS" "Found tool: $tool" ((found_tools++)) else print_status "WARNING" "Tool not found: $tool" fi fi done print_status "INFO" "Found $found_tools/${#expected_tools[@]} expected tools" if [ $found_tools -ge 8 ]; then print_status "SUCCESS" "MCP tools validation passed" return 0 else print_status "ERROR" "Too few tools found ($found_tools/${#expected_tools[@]})" return 1 fi else print_status "ERROR" "Failed to retrieve MCP tools list" return 1 fi } # Function to test MCP resources listing test_mcp_resources() { print_status "INFO" "Testing MCP resources..." local resources_output local timeout_cmd="timeout" if command -v gtimeout >/dev/null 2>&1; then timeout_cmd="gtimeout" elif ! command -v timeout >/dev/null 2>&1; then timeout_cmd="" fi if [ -n "$timeout_cmd" ]; then resources_output=$($timeout_cmd $MCP_INSPECTOR_TIMEOUT npx @modelcontextprotocol/inspector --cli node "$MCP_SERVER_PATH" --method resources/list 2>/dev/null) else resources_output=$(npx @modelcontextprotocol/inspector --cli node "$MCP_SERVER_PATH" --method resources/list 2>/dev/null) fi if [ $? -eq 0 ] && [ -n "$resources_output" ]; then print_status "SUCCESS" "MCP resources list retrieved successfully" # Check for expected resource patterns if echo "$resources_output" | grep -q "adr://"; then print_status "SUCCESS" "Found ADR resources" else print_status "WARNING" "No ADR resources found (may be expected if no ADRs exist)" fi return 0 else print_status "WARNING" "Failed to retrieve MCP resources (may not be implemented)" return 0 fi } # Function to test MCP prompts test_mcp_prompts() { print_status "INFO" "Testing MCP prompts..." local prompts_output local timeout_cmd="timeout" if command -v gtimeout >/dev/null 2>&1; then timeout_cmd="gtimeout" elif ! command -v timeout >/dev/null 2>&1; then timeout_cmd="" fi if [ -n "$timeout_cmd" ]; then prompts_output=$($timeout_cmd $MCP_INSPECTOR_TIMEOUT npx @modelcontextprotocol/inspector --cli node "$MCP_SERVER_PATH" --method prompts/list 2>/dev/null) else prompts_output=$(npx @modelcontextprotocol/inspector --cli node "$MCP_SERVER_PATH" --method prompts/list 2>/dev/null) fi if [ $? -eq 0 ] && [ -n "$prompts_output" ]; then print_status "SUCCESS" "MCP prompts list retrieved successfully" # Check for expected prompts local expected_prompts=( "analyze_architecture" "suggest_improvements" "review_decisions" ) local found_prompts=0 for prompt in "${expected_prompts[@]}"; do if echo "$prompts_output" | grep -q "$prompt"; then print_status "SUCCESS" "Found prompt: $prompt" ((found_prompts++)) fi done if [ $found_prompts -gt 0 ]; then print_status "SUCCESS" "MCP prompts validation passed" else print_status "WARNING" "No expected prompts found (may not be implemented)" fi return 0 else print_status "WARNING" "Failed to retrieve MCP prompts (may not be implemented)" return 0 fi } # Function to test a specific MCP tool call test_mcp_tool_call() { print_status "INFO" "Testing MCP tool call..." # Test analyze_project tool with minimal parameters local tool_output local timeout_cmd="timeout" if command -v gtimeout >/dev/null 2>&1; then timeout_cmd="gtimeout" elif ! command -v timeout >/dev/null 2>&1; then timeout_cmd="" fi if [ -n "$timeout_cmd" ]; then tool_output=$($timeout_cmd $MCP_INSPECTOR_TIMEOUT npx @modelcontextprotocol/inspector --cli node "$MCP_SERVER_PATH" --method tools/call --tool-name analyze_project_ecosystem --tool-arg projectPath="$PROJECT_ROOT" --tool-arg analysisDepth=basic 2>/dev/null) else tool_output=$(npx @modelcontextprotocol/inspector --cli node "$MCP_SERVER_PATH" --method tools/call --tool-name analyze_project_ecosystem --tool-arg projectPath="$PROJECT_ROOT" --tool-arg analysisDepth=basic 2>/dev/null) fi if [ $? -eq 0 ] && [ -n "$tool_output" ]; then print_status "SUCCESS" "MCP tool call executed successfully" # Check if output contains expected content patterns and proper MCP response structure if command -v jq >/dev/null 2>&1; then # Use jq for proper JSON parsing if echo "$tool_output" | jq -e '.content[]? | select(.type == "text")' >/dev/null 2>&1; then local text_content=$(echo "$tool_output" | jq -r '.content[]? | select(.type == "text") | .text' 2>/dev/null) if echo "$text_content" | grep -q -i -E "(analysis|project|ecosystem)"; then print_status "SUCCESS" "Tool output contains valid MCP content with expected keywords" return 0 else print_status "SUCCESS" "Tool output contains valid MCP content structure" return 0 fi else print_status "WARNING" "Tool output does not match expected MCP content structure" return 1 fi else # Fallback to grep-based validation if jq is not available if echo "$tool_output" | grep -q '"content"' && echo "$tool_output" | grep -q '"type":' && echo "$tool_output" | grep -q -i -E "(analysis|project|ecosystem)"; then print_status "SUCCESS" "Tool output contains expected MCP content structure" return 0 elif echo "$tool_output" | grep -q -i -E "(content|analysis|project)"; then print_status "SUCCESS" "Tool output contains expected content keywords" return 0 else print_status "WARNING" "Tool output may be unexpected" echo "DEBUG: First 200 chars of output: $(echo "$tool_output" | head -c 200)..." return 1 fi fi else print_status "ERROR" "Failed to execute MCP tool call" return 1 fi } # Function to generate MCP test report generate_test_report() { local total_tests=$1 local passed_tests=$2 local test_results_file="$PROJECT_ROOT/mcp-test-results.json" print_status "INFO" "Generating MCP test report..." cat > "$test_results_file" << EOF { "timestamp": "$(date -u +"%Y-%m-%dT%H:%M:%SZ")", "mcp_server_path": "$MCP_SERVER_PATH", "node_version": "$(node --version)", "total_tests": $total_tests, "passed_tests": $passed_tests, "success_rate": $(echo "scale=2; $passed_tests * 100 / $total_tests" | bc -l), "status": "$([ $passed_tests -eq $total_tests ] && echo "PASS" || echo "PARTIAL")" } EOF print_status "SUCCESS" "Test report generated: $test_results_file" } # Main execution main() { local total_tests=0 local passed_tests=0 print_status "INFO" "Starting MCP functionality tests..." # Test 1: Prerequisites ((total_tests++)) if check_prerequisites; then ((passed_tests++)) fi # Test 2: Server Health ((total_tests++)) if test_server_health; then ((passed_tests++)) fi # Test 3: Tools List ((total_tests++)) if test_mcp_tools_list; then ((passed_tests++)) fi # Test 4: Resources ((total_tests++)) if test_mcp_resources; then ((passed_tests++)) fi # Test 5: Prompts ((total_tests++)) if test_mcp_prompts; then ((passed_tests++)) fi # Test 6: Tool Call ((total_tests++)) if test_mcp_tool_call; then ((passed_tests++)) fi # Generate report generate_test_report $total_tests $passed_tests # Final results echo print_status "INFO" "MCP Functionality Test Results:" print_status "INFO" "================================" print_status "INFO" "Total Tests: $total_tests" print_status "INFO" "Passed Tests: $passed_tests" print_status "INFO" "Success Rate: $(echo "scale=1; $passed_tests * 100 / $total_tests" | bc -l)%" if [ $passed_tests -eq $total_tests ]; then print_status "SUCCESS" "All MCP functionality tests passed! 🎉" exit 0 elif [ $passed_tests -ge $((total_tests * 2 / 3)) ]; then print_status "WARNING" "Most MCP functionality tests passed (${passed_tests}/${total_tests})" exit 0 else print_status "ERROR" "Too many MCP functionality tests failed (${passed_tests}/${total_tests})" exit 1 fi } # Handle script arguments case "${1:-main}" in "help") echo "Usage: $0 [main|help]" echo " main: Run all MCP functionality tests (default)" echo " help: Show this help message" exit 0 ;; "main"|"") main ;; *) print_status "ERROR" "Unknown command: $1" print_status "INFO" "Use '$0 help' for usage information" exit 1 ;; esac