Skip to main content
Back to challenges
Challenge

Crypto Compliance Agent: Detecting Anomalous Transactions with Gemini 3 Pro Deep Think

This challenge requires you to build a sophisticated compliance monitoring agent using LangGraph. Your system will employ a graph-based workflow to analyze a stream of simulated cryptocurrency transactions, integrating MCP-enabled tools for external blockchain data lookups and historical pattern analysis. Gemini 3 Pro, operating in Deep Think mode, will be crucial for performing hybrid reasoning to interpret complex transaction sequences and flag potentially illicit activities that evade simpler rule-based detection.

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.

Develop a LangGraph-based system using Gemini 3 Pro Deep Think and MCP tool integration to detect anomalous cryptocurrency transactions.

Delivery guide

How work is evaluated

Evaluation

The evaluation will assess the LangGraph system's ability to accurately detect a predefined set of anomalous transaction patterns within a simulated stream, along with the quality of its explanations.

Datasets

Challenge datasets

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

Learning goals

What you should walk away with

  • Design a stateful, graph-based agent workflow using LangGraph.

  • Integrate Gemini 2.5 Pro's Deep Think mode for complex pattern recognition.

  • Implement MCP-enabled tools for real-time blockchain data access and historical analysis.

  • Generate alerts and detailed explanations for suspicious activities.

  • Develop a hybrid reasoning system to identify multi-step anomalous transaction patterns.

Resources and assets

Reference links and supporting material

Dataset notes

Sample data for 1 tasks

Challenge setup

The evaluation will assess the LangGraph system's ability to accurately detect a predefined set of anomalous transaction patterns within a simulated stream, along with the quality of its explanations.

Challenge input

{'transactions': [{'id': 'tx1', 'from': 'addr_a', 'to': 'addr_b', 'amount': 100, 'timestamp': '...', 'type': 'transfer'}]}

Agent execution

No agent tool is listed for this challenge.

Evaluated output

{'anomalies': [{'tx_id': 'tx5', 'reason': 'Circular transaction pattern detected', 'severity': 'high', 'confidence': 0.95}]}

Evaluation checks
  • Each detected anomaly must include 'tx_id', 'reason', 'severity', 'confidence', and 'chain_of_thought'.
  • The system must achieve a minimum recall of 75% on the provided anomalous transaction patterns.
Configured evaluation evidence
  • Precision target: 0.85
Configured runtime
  • Python execution harness

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

No tool records are attached.

Evaluation contract

  • Each detected anomaly must include 'tx_id', 'reason', 'severity', 'confidence', and 'chain_of_thought'.
  • The system must achieve a minimum recall of 75% on the provided anomalous transaction patterns.

Metrics

  • Precision · Target 0.85
  • ReasoningDepthScore · Target 4

Workflow tasks

  1. Task 1

    Input: {'transactions': [{'id': 'tx1', 'from': 'addr_a', 'to': 'addr_b', 'amount': 100, 'timestamp': '...', 'type': 'transfer'}]}

    Output: {'anomalies': [{'tx_id': 'tx5', 'reason': 'Circular transaction pattern detected', 'severity': 'high', 'confidence': 0.95}]}

    Execution harness: Python

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Crypto Compliance Agent: Detecting Anomalous Transactions with Gemini 3 Pro Deep Think