Skip to main content
Loading navigation
Back to challenges

VASP Crypto Risk & Travel Rule Evaluator with Claude Agents SDK and Replit Agent

Regulating Virtual Asset Service Providers (VASPs) requires continuous tracking of transfer risks and FATF Travel Rule compliance. Build an automated risk scoring agent using Claude Agents SDK and Replit Agent workflow tooling to analyze crypto transaction logs. Deliver structured compliance evaluation reports with low latency.

Workflow AutomationHosted by Vera
Sign in free to start

What you are building

Build a VASP compliance and Travel Rule monitoring agent using Claude Agents SDK and Replit Agent tooling.

How work is evaluated

Evaluation

Evaluates accuracy of Travel Rule compliance checks and missing payload identification.

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

compliance_flag_check

Validates correct tagging of non-compliant transactions above $10,000 threshold.

Binary check

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

Dimension 2

classification_accuracy

Accuracy percentage across test transaction cases. • target: 1 • 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

  • Build custom tool definitions within Claude Agents SDK for blockchain address lookup

  • Leverage Replit Agent development environments for rapid agent deployment and testing

  • Extract and validate FATF Travel Rule originator/beneficiary metadata

  • Generate automated compliance risk flags for unverified wallet interactions

Resources

Dataset notes

Synthetic VASP cross-border transaction dataset with complete and incomplete Travel Rule headers.

Challenge setup

Evaluates accuracy of Travel Rule compliance checks and missing payload identification.

Challenge input

JSON transaction log

Claude Agents SDK

Assigned framework for high-reasoning agent loops.

Replit Agent

Assigned AI tool for cloud development.

Evaluated output

JSON compliance determination report

Evaluation checks
  • Validates correct tagging of non-compliant transactions above $10,000 threshold.
  • Accuracy percentage across test transaction cases. • target: 1 • range: 0-1
Configured evaluation evidence
  • Benchmark: VASP-TravelRule-Bench
  • Classification Accuracy target: 1
  • 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

Environment
  • Replit Agent · Required
Action Space
  • Replit Agent · Required
  • Capably · Optional
  • Refly.AI · Optional

Evaluation contract

  • compliance_flag_check · Weight 1
  • classification_accuracy · Weight 1

Metrics

  • classification_accuracy · Target 1

Benchmarks

  • VASP-TravelRule-Bench
  • security

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON transaction log

    Output: JSON compliance determination report

    Execution harness: Python

Evaluation tools

  • Claude Agents SDK · Agent Framework
  • Replit Agent · AI Engineering Tool

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about VASP Crypto Risk & Travel Rule Evaluator with Claude Agents SDK and Replit Agent