socraticode-mcp

What This Skill Does

The socraticode-mcp skill acts as an intelligent bridge between OpenClaw and your local codebase. It utilizes the Model Context Protocol (MCP) to provide high-fidelity, AST-aware indexing and semantic search capabilities. By offloading vector database management to Qdrant, it allows OpenClaw to traverse complex repositories with deep contextual awareness, understanding relationships between functions, classes, and file dependencies. It provides advanced code intelligence, including hybrid search (combining semantic and keyword analysis) and dependency graphing, which ensures that the agent understands how code components are linked together. It is designed to scale with your project, offering live updates via a background file watcher so your codebase index remains relevant as you write code.

Installation

1. Prerequisites: Ensure Docker is running to host the Qdrant container and that Node.js 18+ is installed to support npx and mcporter.
2. MCP Setup: Install the manager by running npm install -g mcporter. Configure your global server list at ~/.openclaw/workspace/config/mcporter.json to include the socraticode command using npx -y socraticode.
3. Database: Initialize the vector engine by running the official Qdrant container: docker run -d --name socraticode-qdrant -p 16333:6333 qdrant/qdrant:v1.17.0.
4. Per-Project Configuration: For specific repositories, create a local mcporter.json file in your project root and invoke commands using the --config mcporter.json flag to maintain isolated, context-specific indices.

Use Cases

• Deep Code Exploration: When onboarding to a new codebase, use codebase_search to find relevant definitions without manual grep commands.
• Dependency Analysis: Utilize codebase_graph_query to identify all files importing a specific module, aiding in impact analysis before refactoring.
• Circular Dependency Detection: Run codebase_graph_circular to catch architectural debt before it leads to runtime issues.
• Documentation Generation: Automatically visualize codebase structure using codebase_graph_visualize for better team communication.

Example Prompts

1. "OpenClaw, please search the codebase for the authentication middleware and explain how it interacts with the user session manager."
2. "Identify any circular dependencies within the utility directory and provide a Mermaid diagram for the modules that are causing the conflict."
3. "Check the indexing status of my current project and start an index update if the last sync was more than ten minutes ago."

Tips & Limitations

• Performance: For large projects, ensure you are using a performant local embedding provider like Ollama; otherwise, utilize OpenAI's API for faster semantic processing.
• Index Freshness: Always run codebase_status before deep analysis to ensure your vector store is synchronized with your disk changes.
• Isolation: Always define a project-specific mcporter.json to prevent pollution of indices across multiple repositories you might be working on simultaneously.