Skillget

agentic-fraud-analysis

mcp
Security Audit
Warn
Health Warn
  • No license — Repository has no license file
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Low visibility — Only 6 GitHub stars
Code Pass
  • Code scan — Scanned 12 files during light audit, no dangerous patterns found
Permissions Pass
  • Permissions — No dangerous permissions requested

No AI report is available for this listing yet.

SUMMARY

An intelligent, multi-agent fraud management system using MCP

README.md

Agentic Fraud Analysis

An intelligent multi-agent system for automated fraud detection, investigation, and mitigation using the Model Context Protocol (MCP).

Architecture

System Architecture

Overview

This system uses a three-agent architecture to handle fraud incidents end-to-end:

Agent Role What it does
Alert Triage Agent First responder Consolidates alerts, correlates patterns, creates/merges incidents
Diagnose Agent Investigator Performs root cause analysis (RCA), identifies attack patterns
Mitigation Agent Problem solver Develops and deploys automated solutions

How It Works

  1. Monitoring Services detect anomalies and send alerts
  2. Alert Triage Agent processes incoming alerts and creates incidents
  3. Diagnose Agent investigates the root cause using available MCP tools
  4. Mitigation Agent proposes and executes fixes
  5. Human-in-the-loop reviews and approves plans via the UI

Key Components

  • AI Proxy: Routes requests through registry, handles authentication
  • MCP Servers: Modular tool servers for specific capabilities (data analysis, ML models, etc.)
  • UI: Dashboard for viewing events and approving agent plans

Getting Started

Prerequisites

  • Python 3.8+
  • uv package manager

Installation

# Clone the repository
git clone https://github.com/yanfeid/agentic-fraud-analysis.git
cd agentic-fraud-analysis

# Install uv if needed
pip install uv

# Set up environment
cp .env.example .env
# Edit .env with your credentials (Azure API keys, etc.)

Running the System

# 1. Start MCP servers (demo mode)
uv run run_toy_mcp.py

# 2. Start the API server
uv run run_api.py

# 3. Start the UI
uv run run_ui.py

The UI will be available at http://localhost:8501

Configuration

Edit .env to configure:

Variable Description
MODEL_NAME LLM model to use (default: gpt-4o)
AZURE_CLIENT_ID Azure credentials for LLM API
GITHUB_TOKEN GitHub access token

License

This project is patent-pending. Commercial use is prohibited.

Reviews (0)

No results found