MCP Prompts Server

# Quick Start Guide ## Using the Skill Scripts ### Quick Reference Paths **Claude Desktop:** ```bash SKILL_DIR="/home/sparrow/.local/share/claude-skills/dev-intelligence-orchestrator/scripts" ``` **Cursor:** ```bash SKILL_DIR="/home/sparrow/.cursor/skills/dev-intelligence-orchestrator/scripts" ``` ### Common Usage Examples #### 1. Detect Project Type ```bash $SKILL_DIR/detect_project_type.sh /path/to/project ``` #### 2. Analyze C++ Code ```bash # First run - captures learning $SKILL_DIR/analyze_cpp.sh src/main.cpp memory . # Second run - uses learned config $SKILL_DIR/analyze_cpp.sh src/main.cpp memory . ``` #### 3. Analyze Python Code ```bash $SKILL_DIR/analyze_python.sh src/utils.py general . ``` #### 4. Parse Build Errors ```bash $SKILL_DIR/parse_build_errors.py build.log esp32 platformio ``` #### 5. Run Tests ```bash $SKILL_DIR/run_tests.sh . tests/ true ``` ### Integration with Claude/Cursor The scripts can be called directly from Claude or Cursor: ``` Use the dev-intelligence-orchestrator skill to analyze the code in src/main.cpp ``` The AI will automatically use the appropriate script from the deployed location. ### Learning Loop Status The scripts will automatically: - 🔍 Query mcp-prompts for learned configurations - ✓ Use learned configs when available - 💡 Capture successful configurations - 📈 Increase confidence with validation ### Prerequisites - mcp-prompts server running (optional, graceful degradation) - Analysis tools: pylint, cppcheck, pytest (as needed) - jq for JSON parsing