AskMeMCP

# AskMe Debug Tools Debug tools for rapid UI development of the AskMe MCP server without requiring an MCP client. ## Overview These tools simulate the behavior of each MCP tool by starting a mock server on port 3000, sending realistic test data, and outputting the formatted result that would be sent to the LLM. ## Available Tools ### 1. Ask One Question (`ask-one-question`) Simulates a free-form question with realistic context about API architecture decisions. ### 2. Ask Multiple Choice (`ask-multiple-choice`) Simulates multiple choice questions with priority ratings and comments for feature prioritization. ### 3. Challenge Hypothesis (`challenge-hypothesis`) Simulates hypothesis evaluation with agreement levels (1-7 scale) for architectural decisions. ### 4. Choose Next (`choose-next`) Simulates decision workflows with multiple strategic options for sprint planning. ## Usage ### Build First ```bash npx nx build askme-debug-tools ``` ### Run Individual Tools ```bash # Using Nx targets (recommended) npx nx debug:ask-one-question askme-debug-tools npx nx debug:ask-multiple-choice askme-debug-tools npx nx debug:challenge-hypothesis askme-debug-tools npx nx debug:choose-next askme-debug-tools # Or using the main CLI directly node dist/askme-debug-tools/main.js ask-one-question node dist/askme-debug-tools/main.js ask-multiple-choice node dist/askme-debug-tools/main.js challenge-hypothesis node dist/askme-debug-tools/main.js choose-next ``` ### Interactive Development ```bash # Terminal 1: Start debug tool npx nx debug:ask-one-question askme-debug-tools # Terminal 2: Start UI development server npx nx serve askme-ui # Browser: Open http://localhost:3000 to interact with mock data ``` ## How It Works 1. **Mock Server**: Each script starts an HTTP server on port 3000 with browser bridge endpoints 2. **Test Data**: Realistic mock data is sent via Server-Sent Events (SSE) 3. **UI Interaction**: Open http://localhost:3000 to see and interact with the UI 4. **Auto Response**: If no UI response within timeout, uses mock human response 5. **LLM Output**: Formatted result is output to stdout, showing what the LLM would receive 6. **Clean Exit**: Server stops and process exits cleanly ## Mock Data Examples ### Ask One Question - **Topic**: API architecture decision for authentication system - **Context**: Microservices platform with specific requirements - **Options**: Auth0/Okta, AWS Cognito, Custom JWT, Keycloak ### Ask Multiple Choice - **Questions**: Feature prioritization, deployment environments, testing strategies - **Format**: Priority ratings (-3 to +3) with detailed comments ### Challenge Hypothesis - **Topic**: Monolith to microservices migration decision - **Format**: 7-point agreement scale with detailed reasoning ### Choose Next - **Topic**: Q4 sprint planning with 5 strategic options - **Format**: Detailed options with icons, descriptions, and strategic context ## Development Benefits - **Fast Iteration**: No need to set up MCP client connections - **Realistic Data**: Production-quality test scenarios - **UI Focus**: Test UI behavior with consistent, predictable data - **Debug Friendly**: Easy to modify mock responses and timing - **Standalone**: Works independently of the main MCP server ## File Structure ``` askme-debug-tools/src/ ├── debug-ask-one-question.ts # Ask one question simulator ├── debug-ask-multiple-choice.ts # Multiple choice simulator ├── debug-challenge-hypothesis.ts # Hypothesis evaluation simulator ├── debug-choose-next.ts # Decision workflow simulator ├── main.ts # CLI dispatcher └── shared/ ├── mock-server.ts # HTTP server with SSE endpoints ├── mock-data.ts # Realistic test data └── response-simulator.ts # Auto-response simulation ``` ## Port Usage - **Port 3000**: Mock server (matches real askme-server default) - **Port 4200**: UI development server (Angular default) ## Tips - Run the debug tool first, then start UI development server - Use different terminals to see both debug output and UI logs - Modify mock data in `shared/mock-data.ts` for different test scenarios - Adjust response delays in individual debug scripts - Check stdout for formatted LLM results, stderr for debug logs