Skip to main content
Loading navigation
Back to challenges

Build a Stablecoin Settlement Compliance Agent with CrewAI and OpenAI o3

Cross-border stablecoin settlements require automated risk compliance, address sanction screening, and multi-currency liquidity validation. Build a CrewAI agentic workflow using OpenAI o3 to parse transaction events, verify compliance rules, and flag high-risk transfers in real time while reducing false positives below 5%.

CybersecurityHosted by Vera
Sign in free to start

What you are building

Orchestrate multi-agent role-playing teams using CrewAI and OpenAI o3 reasoning to evaluate stablecoin ledger transactions against OFAC sanctions and velocity thresholds.

How work is evaluated

Evaluation

Evaluates the agentic workflow on classifying real-time stablecoin transactions into compliant, flagged, or rejected categories.

Datasets

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

How submissions are scored

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

schema_validation_test

Ensures output conforms strictly to the expected decision schema

Binary check

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

Dimension 2

classification_accuracy

Percentage of test cases correctly assigned compliance decisions • target: 0.95 • 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

  • Construct a multi-agent CrewAI orchestration flow for cross-border stablecoin compliance

  • Integrate OpenAI o3 reasoning models for complex compliance rule validation

  • Implement real-time address risk scoring and sanction list matching

  • Establish automated audit logging and investigator triage queue generation

Resources

Dataset notes

Synthetic dataset of 500 stablecoin transaction payloads with annotated risk levels and sanction list correlations.

Challenge setup

Evaluates the agentic workflow on classifying real-time stablecoin transactions into compliant, flagged, or rejected categories.

Challenge input

JSON transaction event containing sender, recipient, amount, token, and chain metadata

CrewAI

Provides multi-agent task execution and role specialization.

OpenAI o3

Delivers superior logical chain-of-thought processing for complex rule evaluations.

Evaluated output

JSON containing risk_score, compliance_decision, matched_rules, and explanation

Evaluation checks
  • Ensures output conforms strictly to the expected decision schema
  • Percentage of test cases correctly assigned compliance decisions • target: 0.95 • range: 0-1
Configured evaluation evidence
  • Benchmark: Financial Crime Detection Benchmark v1
  • Classification Accuracy target: 0.95
  • 1 public reference case
Configured runtime
  • Python execution harness
  • Python sandbox (unavailable on Versalist)

Shows the first task, up to three tools, and up to three evidence items.

View technical recipe

Configured tools

Action Space
  • CrewAI · Required
  • OpenAI · Optional
Policy Serving
  • o3 · Optional

Evaluation contract

  • schema_validation_test · Weight 1
  • classification_accuracy · Weight 1

Metrics

  • classification_accuracy · Target 0.95

Benchmarks

  • Financial Crime Detection Benchmark v1
  • security

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON transaction event containing sender, recipient, amount, token, and chain metadata

    Output: JSON containing risk_score, compliance_decision, matched_rules, and explanation

    Execution harness: Python

Evaluation tools

  • CrewAI · Agent Orchestration Framework
  • OpenAI o3 · Reasoning LLM Engine

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Build a Stablecoin Settlement Compliance Agent with CrewAI and OpenAI o3