Skip to main content
Back to challenges
Challenge

Anti-Scam Mule Account Pattern Detection Engine with Mastra AI

Mitigating public scam networks requires rapid ingestion and analysis of reported mule accounts and transaction flows. Construct an automated scam analysis workflow using Mastra AI in TypeScript to trace scam signals across transaction logs. Deliver rapid anomaly cluster tagging and automated emergency hold recommendations.

Workflow AutomationHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Build an anti-scam telemetry parsing and mule account detection engine using Mastra AI framework.

Delivery guide

How work is evaluated

Evaluation

Evaluates detection accuracy for mule accounts and speed of triggering protective holds.

Datasets

Challenge datasets

Public datasets, plus the private datasets you have access to.

Evaluation rubric

How submissions are scored

These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

freeze_recommendation_check

Validates that accounts with high mule probability trigger emergency freeze.

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

mule_detection_f1

F1 score on classifying fraudulent mule accounts. • target: 0.96 • range: 0-1

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

What you should walk away with

  • Build Mastra AI workflows for multi-step scam telemetry ingestion and processing

  • Implement built-in Mastra RAG tools to cross-reference known scam phone numbers and accounts

  • Design dynamic action graphs to trigger automated bank hold alerts

  • Deploy stateful Mastra agents with persistent session memory

Resources and assets

Reference links and supporting material

Dataset notes

Synthetic scam report feeds and corresponding mule account velocity logs.

Challenge setup

Evaluates detection accuracy for mule accounts and speed of triggering protective holds.

Challenge input

JSON report containing account activity

Mastra AI

Assigned framework for TypeScript workflow orchestration.

Evaluated output

JSON threat evaluation and action recommendation

Evaluation checks
  • Validates that accounts with high mule probability trigger emergency freeze.
  • F1 score on classifying fraudulent mule accounts. • target: 0.96 • range: 0-1
Configured evaluation evidence
  • Benchmark: AntiScam-Mule-Bench
  • Mule Detection F1 target: 0.96
  • 1 public reference case
Configured runtime
  • JavaScript execution harness
  • javascript sandbox (unavailable on Versalist)

Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.

View technical recipe

Configured tools

Action Space
  • Mastra AI · Required
  • Merge · Optional
Observation
  • Helicone · Optional
Orchestration
  • Mastra AI · Required

Evaluation contract

  • freeze_recommendation_check · Weight 1
  • mule_detection_f1 · Weight 1

Metrics

  • mule_detection_f1 · Target 0.96

Benchmarks

  • AntiScam-Mule-Bench
  • security

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON report containing account activity

    Output: JSON threat evaluation and action recommendation

    Execution harness: JavaScript

Evaluation tools

  • Mastra AI · Workflow Framework

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Anti-Scam Mule Account Pattern Detection Engine with Mastra AI