Pathscan MCP Server

🚀 Usage Guide for MCP Service

This guide provides a step-by-step process to set up and use the MCP service, including installing dependencies, deploying related tools, and configuring the service for use.

🚀 Quick Start

Step 1: Install Dependencies for dirsearch

Navigate to the dirsearch directory and execute the following command to install dependencies (here, we choose to install them globally for simplicity):

pip install -r .\requirements.txt

Step 2: Install an Additional Package in the Global Environment

pip install setuptools

Step 3: Install the uv Manager

uv is an extremely fast Python package and project manager written in Rust. Use the following command to install it:

powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

Step 4: Install Dependencies for the MCP Service

uv sync

Step 5: Deploy firecrawl Locally

firecrawl is an AI crawler tool with data cleaning capabilities and supports local Docker deployment.

Reference Articles:

• https://www.cnblogs.com/skystrive/p/18893148
• https://docs.firecrawl.dev/contributing/guide

Troubleshooting: When using the locally deployed firecrawl for web crawling and getting an Unauthorized error, set the value of USE_DB_AUTHENTICATION in the .env configuration file to false. You can directly use the provided .env file.

To test if your firecrawl is running, open the URL: http://{firecrawl server IP}:3002/test. If you see Hello, world!, the service is running properly.

Step 6: Edit Your config.py File

• FIRECRAWL_HOST is the host address where your firecrawl is running.
• GLOBAL_PYTHON_PATH is the global Python address. You can get it by entering where python in the cmd.

Step 7: Start Your MCP Service

uv run main.py

If you see the following logs, it means your service has started successfully.

Step 8: Add This Service to cline

Edit the configuration file directly:

{
    "mcpServers": {
        "path scanner": {
            "url": "http://127.0.0.1:8000/sse",
            "disabled": false,
            "autoApprove": [],
            "timeout": 1800
        }
    }
}

This indicates that your MCP server can be used by cline.

💻 Usage Examples

Prompt

You are a website security assistant. Please call the existing MCP service to scan the specified website and return the structured results (in tabular form). Based on the MCP scan results:

• Extract and summarize the technology stack used by the website.
• Classify and organize the vulnerability information of the website according to the risk levels in the scan report (High Risk/Medium Risk/Low Risk).
• For the URLs marked as high-risk vulnerabilities in the report, further read the page content of these URLs and generate content summaries. Finally, please summarize the data into the following structure:
• Technology Stack
• Vulnerability Information (Classified by Severity)
• High-Risk Vulnerability Related URLs and Content Summaries Target Website URL:

📸 Effect Display

📅 TODO

When the server receives a client request before it is fully initialized, the solution is to initialize the service before using it formally.

🔗 Reference

This project is inspired by the project ai_dirscan.