MCP Servers

模型上下文协议服务器、框架、SDK 和模板的综合目录。

I
Incidentflow Mcpserver

MCP server by taoufikmohamed

创建于 2/2/2026
更新于 23 days ago
Repository documentation and setup instructions

IncidentFlow

Automated Incident Response Platform for Windows

IncidentFlow monitors Windows Event Logs, uses AI to classify the severity of incidents, and automatically reports them to Slack. It runs as a set of background Windows Services for continuous operation.

🚀 Features

  • Real-time Monitoring: Detects errors from Windows Event Logs instantly.
  • AI-Powered Severity: Uses DeepSeek AI to intelligently classify incidents as CRITICAL, HIGH, MEDIUM, or LOW.
  • Slack Integration: Sends formatted alerts directly to your Slack workspace.
  • Resilient Architecture: Runs as three decoupled microservices (Log Agent, MCP Server, Slack Agent) managed by NSSM.

🛠️ Installation

Prerequisites

  • Python 3.11+ installed and added to PATH.
  • NSSM (included or installed via Chocolatey/Scoop).
  • Administrator Privileges (required to install services).

Quick Start

We provide an automated PowerShell installer to set up everything for you.

  1. Clone the repository:

    git clone https://github.com/your-repo/incidentflow.git
    cd incidentflow
    
  2. Run the Installer (as Administrator):

    .\install\install_services.ps1
    
    • The script will ask for your API keys (INCIDENTFLOW_API_KEY, DEESEEK_API_KEY, SLACK_WEBHOOK_URL) if they are not already set.
    • It will install and start all three services automatically.

⚙️ Configuration

The system uses the following environment variables (set automatically by the installer):

| Variable | Description | | :--- | :--- | | INCIDENTFLOW_API_KEY | Secure key for internal API communication. | | DEESEEK_API_KEY | API Key for DeepSeek AI (for severity classification). | | SLACK_WEBHOOK_URL | Webhook URL for your Slack channel. |


🧪 Testing

You can verify the system is working by sending a manual test incident.

Run the Test Script

.\test_flow.ps1

This script sends a simulated "CRITICAL" incident to the MCP server, which should then appear in your Slack.

Test Severity Classification

To test how the AI classifies different types of incidents:

.\test_severity.ps1

🔍 Troubleshooting

Check Service Status

nssm status IncidentFlow-MCP
nssm status IncidentFlow-Slack
nssm status IncidentFlow-LogAgent

View Logs

Logs are located in the logs/ directory.

  • MCP Server Errors: logs/IncidentFlow-MCP.err
  • Slack Agent Errors: logs/IncidentFlow-Slack.err

To tail the logs in real-time:

Get-Content logs\IncidentFlow-MCP.err -Wait

Restart Services

If you need to apply changes or restart the system:

nssm restart IncidentFlow-MCP
nssm restart IncidentFlow-Slack
nssm restart IncidentFlow-LogAgent

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

快速设置
此服务器的安装指南

安装包 (如果需要)

uvx incidentflow-mcpserver

Cursor 配置 (mcp.json)

{ "mcpServers": { "taoufikmohamed-incidentflow-mcpserver": { "command": "uvx", "args": [ "incidentflow-mcpserver" ] } } }