Skip to main content
Back to challenges
Challenge

Develop an AutoGen Account Aggregator Credit Verification Agent in E2B

Financial service providers require isolated environments to evaluate sensitive user financial statements without exposing raw PII. Build a multi-agent conversational audit system using AutoGen inside an E2B sandbox environment to compute underwriting scores while maintaining 100% consent ledger compliance.

Business OperationsHosted by Vera
Sign in free to start
Challenge brief

What you are building

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

Construct an AutoGen multi-agent system executing code safely inside an E2B cloud sandbox to compute Account Aggregator financial scores.

Delivery guide

How work is evaluated

Evaluation

Evaluates multi-agent score accuracy and sandbox isolation security during financial statement evaluation.

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

consent_check

Confirms valid consent verification before calculation

Binary check

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

Dimension 2

calculated_score

Underwriting credit risk score output • target: 780 • range: 300-900

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

  • Configure AutoGen UserProxyAgent and AssistantAgent pairs for financial data processing

  • Deploy E2B Code Interpreter micro-sandboxes for safe execution of financial metric scripts

  • Validate consent artifact tokens issued by Account Aggregator frameworks

  • Generate comprehensive debt-to-income and cash flow risk reports

Resources and assets

Reference links and supporting material

Dataset notes

Synthetic Account Aggregator cash flow payloads and financial statement metrics.

Challenge setup

Evaluates multi-agent score accuracy and sandbox isolation security during financial statement evaluation.

Challenge input

JSON payload containing consent_token and statement data

AutoGen

Multi-agent conversation orchestration

E2B

Secure code execution in isolated environments

Evaluated output

JSON report containing credit_score and risk_tier

Evaluation checks
  • Confirms valid consent verification before calculation
  • Underwriting credit risk score output • target: 780 • range: 300-900
Configured evaluation evidence
  • Benchmark: AA Financial Underwriting Benchmark
  • Calculated Score target: 780
  • 1 public reference case
Configured runtime
  • Python execution harness
  • Python 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

Environment
  • E2B · Optional
Action Space
  • AutoGen · Required
  • E2B · Optional
  • Arcee AI · Optional

Evaluation contract

  • consent_check · Weight 1
  • calculated_score · Weight 1

Metrics

  • calculated_score · Target 780

Benchmarks

  • AA Financial Underwriting Benchmark
  • security

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON payload containing consent_token and statement data

    Output: JSON report containing credit_score and risk_tier

    Execution harness: Python

Evaluation tools

  • AutoGen · Agent Framework
  • E2B · Sandbox Execution

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Develop an AutoGen Account Aggregator Credit Verification Agent in E2B