Gemini MCP

Instructions

Implementation Reference

• src/tools/enhanced-tools.js:553-688 (handler)

  The main execution handler for the 'mcp__gemini__debug_master' tool. It processes error messages, code, stack traces, and context to provide root cause analysis, optional execution simulation, fix suggestions, and validation guidance using AI calls.

      handler: async (args) => {
        const { error_message, code, stack_trace, context = '', language = 'javascript', simulate_execution = true } = args;
        validateString(error_message, 'error_message');
        
        const timer = performanceMonitor.startTimer('debug_master');
        
        let debugPrompt = `Advanced debugging analysis for ${language} error:
  
  **Error**: ${error_message}
  
  ${stack_trace ? `**Stack Trace**:\n\`\`\`\n${stack_trace}\n\`\`\`\n` : ''}
  
  ${code ? `**Code Context**:\n\`\`\`${language}\n${code}\n\`\`\`\n` : ''}
  
  ${context ? `**Additional Context**: ${context}\n` : ''}
  
  Provide comprehensive debugging analysis:
  
  1. **Root Cause Analysis**
     - Primary cause identification
     - Contributing factors
     - Error propagation path
  
  2. **Code Flow Analysis**
     - Execution path leading to error
     - Variable states at error point
     - Control flow issues
  
  3. **Debugging Strategy**
     - Step-by-step debugging approach
     - Key checkpoints to examine
     - Data points to verify
  
  4. **Fix Solutions**
     - Primary fix with code example
     - Alternative approaches
     - Prevention strategies
  
  5. **Testing Validation**
     - How to verify the fix
     - Edge cases to test
     - Regression testing points
  
  Be specific and actionable.`;
  
        const analysis = await aiClient.call(debugPrompt, 'debug', { maxTokens: 3000 });
        
        let result = `🐛 **Debug Master Analysis** (${language})
  
  ${analysis}`;
  
        // Execution simulation if enabled and code is provided
        if (simulate_execution && code) {
          const simulationPrompt = `Simulate the execution of this code step by step:
  
  \`\`\`${language}
  ${code}
  \`\`\`
  
  Error: ${error_message}
  
  Provide:
  1. **Step-by-step execution trace**
     - Line-by-line execution flow
     - Variable values at each step
     - Function calls and returns
  
  2. **Error Point Analysis**
     - Exact line where error occurs
     - Variable states at error
     - Memory/scope situation
  
  3. **Execution Scenarios**
     - Normal execution path
     - Error-triggering scenario
     - Edge case behaviors
  
  4. **State Verification**
     - Expected vs actual values
     - Assumption violations
     - Boundary condition issues
  
  Format as a detailed execution trace.`;
  
          const simulation = await aiClient.call(simulationPrompt, 'analysis');
          
          result += `
  
  ---
  
  🔄 **Execution Simulation**
  
  ${simulation}`;
        }
        
        // Generate fix validation
        if (code) {
          const validationPrompt = `Based on the debugging analysis, validate potential fixes:
  
  Original Code:
  \`\`\`${language}
  ${code}
  \`\`\`
  
  Error: ${error_message}
  
  Provide:
  1. **Fix Validation Checklist**
     - Critical points to verify
     - Test cases to validate
     - Edge conditions to check
  
  2. **Regression Risk Assessment**
     - Areas that might be affected
     - Side effects to monitor
     - Backward compatibility concerns
  
  3. **Monitoring Recommendations**
     - What to watch after deployment
     - Logging improvements
     - Alert configurations`;
  
          const validation = await aiClient.call(validationPrompt, 'review');
          
          result += `
  
  ---
  
  ✅ **Fix Validation Guide**
  
  ${validation}`;
        }
        
        timer.end();
        return result;
      }

• src/tools/enhanced-tools.js:545-552 (schema)

  The input parameters schema defining the expected arguments for the tool, including required error_message and optional fields like code, stack_trace, etc.

  parameters: {
    error_message: { type: 'string', description: 'Error message or description', required: true },
    code: { type: 'string', description: 'Code where error occurs' },
    stack_trace: { type: 'string', description: 'Stack trace if available' },
    context: { type: 'string', description: 'Additional context' },
    language: { type: 'string', description: 'Programming language', default: 'javascript' },
    simulate_execution: { type: 'boolean', description: 'Enable execution simulation', default: true }
  },

• src/tools/registry.js:235-239 (registration)

  The registration code that imports and registers all tools from the enhancedTools module, which includes 'mcp__gemini__debug_master', by calling registerToolsFromModule(enhancedTools).

  this.registerToolsFromModule(codeTools);
  this.registerToolsFromModule(analysisTools);
  this.registerToolsFromModule(enhancedTools);
  this.registerToolsFromModule(businessTools);
  this.registerToolsFromModule(licenseTools);

• src/tools/registry.js:7-7 (registration)

  Import statement that brings in the enhancedTools object containing the definition of 'mcp__gemini__debug_master'.

  import { enhancedTools } from './enhanced-tools.js';