Skip to main content
Glama
mcp-prompts.md9.06 kB
# MCP Prompts Guide CodeRAG includes 4 guided prompts to help you effectively use the tools: ## 1. `analyze_codebase` - Comprehensive Analysis Guide **Purpose:** Get step-by-step guidance for analyzing your codebase **Usage:** ``` Use the analyze_codebase prompt for my TypeScript project ``` **What it provides:** - Instructions for exploring existing graph data - Steps to understand codebase structure - Relationship analysis guidance - Quality assessment workflow **Best for:** - First-time codebase exploration - Comprehensive quality assessment - Understanding unfamiliar codebases - Preparing architecture documentation **Example workflow:** 1. Prompt guides you to scan the project 2. Shows how to get overview with `get_project_summary` 3. Explains finding architectural issues 4. Demonstrates class-level analysis 5. Provides guidance on interpreting results ## 2. `setup_code_graph` - Project Setup Guide **Purpose:** Step-by-step guide to set up a code graph for a new project **Usage:** ``` Use the setup_code_graph prompt for language=java ``` **What it provides:** - Scanning instructions - Validation steps - Structure analysis guidance - Relationship exploration tips **Best for:** - New CodeRAG users - Setting up analysis for new projects - Ensuring proper configuration - Learning the basic workflow **Example workflow:** 1. Guides through initial project scanning 2. Shows how to verify scan results 3. Explains basic node and relationship queries 4. Demonstrates essential analysis tools 5. Provides troubleshooting tips ## 3. `find_dependencies` - Dependency Analysis **Purpose:** Guide to find class dependencies and method calls **Usage:** ``` Use the find_dependencies prompt for target_class=UserService ``` **What it provides:** - Instructions to find what a class depends on - Steps to find what depends on the class - Coupling analysis guidance - Visualization approaches **Best for:** - Analyzing specific classes or components - Understanding coupling and dependencies - Planning refactoring efforts - Impact analysis for changes **Example workflow:** 1. Shows how to find outgoing dependencies 2. Demonstrates finding incoming dependencies 3. Explains coupling metrics interpretation 4. Guides through relationship visualization 5. Provides refactoring recommendations ## 4. `analyze_inheritance` - Inheritance Analysis **Purpose:** Guide to analyze inheritance hierarchies and interface implementations **Usage:** ``` Use the analyze_inheritance prompt for class_or_interface=BaseRepository ``` **What it provides:** - Hierarchy mapping instructions - Implementation discovery steps - Design pattern identification - Architecture evaluation guidance **Best for:** - Object-oriented design analysis - Understanding inheritance structures - Identifying design patterns - Evaluating abstraction levels **Example workflow:** 1. Maps complete inheritance hierarchy 2. Finds all implementations of interfaces 3. Analyzes depth and breadth of inheritance 4. Identifies potential design issues 5. Suggests architectural improvements ## Using Prompts Effectively ### 1. Choose the Right Prompt **Start with `setup_code_graph`** when: - New to CodeRAG - Working with a new project - Need to verify setup **Use `analyze_codebase`** when: - Want comprehensive analysis - Need to understand overall architecture - Preparing quality reports **Use `find_dependencies`** when: - Investigating specific classes - Planning refactoring - Understanding coupling **Use `analyze_inheritance`** when: - Working with OOP codebases - Analyzing design patterns - Evaluating abstraction layers ### 2. Sequential Usage **Recommended Order for New Projects:** 1. `setup_code_graph` - Initial setup and validation 2. `analyze_codebase` - Overall understanding 3. `find_dependencies` - Specific class investigation 4. `analyze_inheritance` - OOP structure analysis ### 3. Interactive Guidance **Each prompt provides:** - Clear step-by-step instructions - Tool recommendations with exact parameters - Interpretation guidance for results - Follow-up suggestions - Troubleshooting tips **Example interaction:** ``` User: Use the analyze_codebase prompt for my TypeScript project Assistant: I'll guide you through a comprehensive codebase analysis... Step 1: First, let's scan your project Use: scan_dir with your project path Step 2: Get project overview Use: get_project_summary Step 3: Identify issues Use: find_architectural_issues [Continues with detailed guidance] ``` ## Prompt Parameters ### Common Parameters **Language-specific prompts:** - `language`: Programming language (typescript, java, python, etc.) - Affects tool recommendations and best practices **Class-specific prompts:** - `target_class`: Specific class to analyze - `class_or_interface`: Class or interface name for inheritance analysis **Project context:** - `project_id`: Specific project in multi-project setup - Automatically inferred if only one project exists ### Example Parameters ``` # Language-specific setup analyze_codebase prompt for language=java # Class-specific analysis find_dependencies prompt for target_class=com.example.UserService # Interface hierarchy analysis analyze_inheritance prompt for class_or_interface=Repository # Multi-project context setup_code_graph prompt for language=typescript project_id=frontend-app ``` ## Advanced Prompt Usage ### 1. Combining Prompts **Sequential Analysis:** ``` 1. setup_code_graph for language=java 2. analyze_codebase for comprehensive review 3. find_dependencies for UserService 4. analyze_inheritance for BaseEntity ``` **Focused Investigation:** ``` 1. find_dependencies for problem class 2. analyze_inheritance for base classes 3. analyze_codebase for quality metrics ``` ### 2. Multi-Project Workflows **Portfolio Analysis:** ``` 1. setup_code_graph for each project 2. analyze_codebase for each project 3. Compare results across projects ``` **Migration Planning:** ``` 1. analyze_codebase for legacy system 2. find_dependencies for core components 3. analyze_inheritance for key abstractions ``` ### 3. Team Collaboration **Code Review Preparation:** ``` 1. analyze_codebase for overview 2. find_dependencies for changed classes 3. Share results with team ``` **Onboarding New Developers:** ``` 1. setup_code_graph for project setup 2. analyze_codebase for architecture understanding 3. analyze_inheritance for design patterns ``` ## Prompt Output and Follow-up ### Understanding Prompt Results **Prompts provide:** - Specific tool calls to make - Expected output explanation - Interpretation guidance - Next step recommendations **Example output structure:** ``` === CodeRAG Guided Analysis === Step 1: Project Scanning > Use tool: scan_dir > Parameters: {"directory_path": "/path/to/project"} > Expected: Project entities and relationships Step 2: Overview Analysis > Use tool: get_project_summary > Expected: Quality metrics and statistics > Look for: High complexity classes, architectural issues [Continues with detailed guidance] ``` ### Follow-up Actions **After using prompts:** 1. Execute recommended tool calls 2. Review and interpret results 3. Follow suggested next steps 4. Use additional prompts as needed 5. Document findings and decisions ## Troubleshooting Prompts ### Common Issues **Prompt doesn't provide expected guidance:** - Verify prompt name and parameters - Check project context (single vs multi-project) - Ensure project is properly scanned **Tools mentioned in prompt not available:** - Verify CodeRAG MCP server is connected - Check tool permissions and configuration - Restart AI assistant if needed **Results don't match prompt expectations:** - Verify project was scanned successfully - Check for scanning errors or warnings - Consider re-scanning with different options ### Getting Help **When prompts aren't helpful:** 1. Check the [Available Tools](available-tools.md) guide 2. Review [Scanner Usage](scanner-usage.md) for setup issues 3. Consult [Troubleshooting](troubleshooting.md) for common problems 4. Try a different prompt that might be more appropriate ## Best Practices ### 1. Prompt Selection - Start with setup prompts for new projects - Use specific prompts for focused analysis - Combine prompts for comprehensive understanding ### 2. Parameter Usage - Provide specific class names when possible - Include language context for better guidance - Use project_id in multi-project environments ### 3. Result Interpretation - Follow prompt guidance for interpreting results - Look for patterns and trends, not just individual metrics - Consider architectural context when evaluating findings ### 4. Documentation - Document key findings from prompt-guided analysis - Share insights with team members - Use results to plan improvements and refactoring ## Next Steps - [Available Tools](available-tools.md) - Detailed reference for all CodeRAG tools - [Quality Metrics](quality-metrics.md) - Understanding analysis results - [Troubleshooting](troubleshooting.md) - Solving common issues

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/JonnoC/CodeRAG'

If you have feedback or need assistance with the MCP directory API, please join our Discord server