fraud-detection

Detect and block fraudulent e-commerce orders with risk scoring and 3DS enforcement.

14|3|Updated Mar 22, 2026
One-click install
npx skills add https://github.com/tomtoto757/ecomm-ai-skills-hub --skill fraud-detection-tomtoto757
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: fraud-detection
Source: https://github.com/tomtoto757/ecomm-ai-skills-hub/tree/main/skills/platform-integrations-infrastructure/finsilabs/security-compliance/fraud-detection
Command: npx skills add https://github.com/tomtoto757/ecomm-ai-skills-hub --skill fraud-detection-tomtoto757

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Protects online stores from fraudulent orders that lead to chargebacks, lost goods, and dispute fees by combining platform-native risk signals, machine-learning scores, 3D Secure enforcement, velocity checks, and human review workflows.

Core Features & Use Cases

  • Platform-native integrations: Guides configuration for Shopify Fraud Analysis, Stripe Radar, Signifyd, NoFraud, and BigCommerce/PayPal/Braintree options.
  • Composite risk scoring: Combines Stripe Radar scores with local signals (velocity violations, address mismatches, device fingerprints, email age) to produce auditable risk scores and decisions.
  • 3DS & payment handling: Enforces 3D Secure challenges for high-risk transactions and recommends authorize-only (manual capture) flows for orders held for review.
  • Velocity checks & Redis: Designs Redis-backed counters and daily spend tracking to block card testing and high-frequency abuse across instances.
  • Manual review queue: Details holding orders, notifying fraud teams via Slack, expiring unreviewed orders after 48 hours, and minimizing client-facing leakage of internal rules.
  • Operational guidance: KPIs to monitor (false positive rate, chargeback rate), whitelisting strategies, and rule-rotation best practices.

Quick Start

Run a review: analyze recent orders, compute composite risk scores using Stripe Radar plus local signals, require 3DS for scores above 70, and place orders scoring 50–79 into an authorize-only manual review queue.

Frequently Asked Questions about fraud-detection

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I prevent chargebacks with 3D Secure and Stripe Radar on Shopify?

To prevent chargebacks, you can enforce 3D Secure challenges for high-risk transactions and combine Stripe Radar scores with platform-native signals like Shopify Fraud Analysis to create a composite risk score.

Can I use Redis for velocity checks to stop card testing attacks?

Yes, you can design Redis-backed counters and daily spend tracking to throttle high-frequency abuse and block card testing across your e-commerce instances.

What is an authorize-only manual review workflow for e-commerce fraud?

An authorize-only manual review workflow holds suspicious orders for human inspection rather than capturing payment immediately, expiring unreviewed orders after 48 hours and notifying fraud teams via Slack.

Does this fraud detection approach work with WooCommerce and BigCommerce?

Yes, this fraud detection approach applies to Shopify, WooCommerce, BigCommerce, and custom headless platforms, integrating platform-native tools and Stripe Radar for risk scoring.

How do I compute composite risk scores for fraudulent orders?

You compute composite risk scores by combining Stripe Radar machine-learning scores with local signals like velocity violations, address mismatches, device fingerprints, and email age to produce auditable decisions.

When should I require 3D Secure for high-risk transactions?

You should require 3D Secure for transactions exceeding a configurable risk score threshold, such as scores above 70, while placing orders scoring 50 to 79 into an authorize-only manual review queue.