How does it help reduce false positives?

It provides principles for tuning thresholds per risk tier and combining rules with ML to ensure high-accuracy scoring that doesn't disrupt legitimate customers.

Can this help with 3DS2 implementation?

Yes, it includes specialized guidance for 3DS2 frictionless versus challenge flows and optimizing implementation via processor SDKs to ensure liability shift.

Does it support real-time monitoring?

Yes, it covers the design of real-time scoring pipelines, feature freshness requirements, and automated decision-making workflows.

What fraud detection models does this skill support?

The skill provides implementation patterns for both rule-based systems (velocity checks, geo-anomalies) and ML models such as Isolation Forests and autoencoders.

Fintech Fraud Detection

Name: Fintech Fraud Detection
Author: rnavarych

byrnavarych

•

Seguridad y Pruebas

Implements advanced fraud prevention systems using rule-based logic, machine learning anomalies, and behavioral biometrics.

This skill provides specialized guidance for building robust fraud detection systems within fintech and e-commerce environments. It covers the implementation of hybrid detection models—combining traditional velocity checks and geographic anomaly rules with sophisticated machine learning techniques like Isolation Forests and autoencoders. Users can leverage this skill to design real-time scoring pipelines, integrate device fingerprinting, manage 3DS2 authentication flows, and establish feedback loops for analyst case management to reduce chargebacks and minimize customer friction.

Características Principales

01Case management workflows and chargeback prevention strategies

02Hybrid rule-based and ML-based anomaly detection systems

0311 GitHub stars

04Real-time fraud scoring pipeline and feature store architecture

053DS2 implementation and frictionless approval optimization

06Device fingerprinting and behavioral biometrics integration

Casos de Uso

01Developing a feedback loop for human analysts to retrain machine learning models

02Integrating identity verification and device risk scoring into authentication flows

03Building a real-time transaction monitoring system for a payment gateway

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add rnavarych/alpha-engineer fraud-detection

For use in Claude.ai and ChatGPT

Características Principales

01Case management workflows and chargeback prevention strategies

02Hybrid rule-based and ML-based anomaly detection systems

0311 GitHub stars

04Real-time fraud scoring pipeline and feature store architecture

053DS2 implementation and frictionless approval optimization

06Device fingerprinting and behavioral biometrics integration

Casos de Uso

01Developing a feedback loop for human analysts to retrain machine learning models

02Integrating identity verification and device risk scoring into authentication flows

03Building a real-time transaction monitoring system for a payment gateway

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add rnavarych/alpha-engineer fraud-detection

For use in Claude.ai and ChatGPT