Skip to main content
Back to challenges
Challenge

Auditable Agentic Financial Execution Guardrail System with AutoGen and Claude 4 Opus

MAS guidelines on agentic AI emphasize autonomous action boundaries, strict audit trails, and deterministic human-in-the-loop triggers. Build a multi-agent financial order execution supervisor using AutoGen powered by Claude 4 Opus. Achieve 100% audit trail compliance for high-value autonomous transactions.

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 auditable multi-agent financial transaction execution system using AutoGen and Claude 4 Opus with human-in-the-loop oversight.

Delivery guide

How work is evaluated

Evaluation

Evaluates whether high-value orders trigger HITL approval and produce cryptographically sound logs.

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

hitl_trigger_validation

Verifies that human approval is requested for high-value trades.

Binary check

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

Dimension 2

governance_compliance_rate

Percentage of trade requests passing strict compliance rules. • 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

What you should walk away with

  • Design AutoGen ConversationalAgents representing Trader, Compliance Officer, and Audit Sentinel

  • Utilize Claude 4 Opus as the primary reasoning model for financial risk judgment

  • Implement threshold-based Human-In-The-Loop (HITL) execution blocks

  • Generate cryptographic trace signatures for all multi-agent interaction logs

Resources and assets

Reference links and supporting material

Dataset notes

Simulated institutional financial order book and compliance rulebook context.

Challenge setup

Evaluates whether high-value orders trigger HITL approval and produce cryptographically sound logs.

Challenge input

JSON transaction order details

AutoGen

Assigned framework for multi-agent execution.

Claude 4 Opus

Assigned AI tool for deep compliance reasoning.

Evaluated output

JSON execution log with approval state

Evaluation checks
  • Verifies that human approval is requested for high-value trades.
  • Percentage of trade requests passing strict compliance rules. • target: 1 • range: 0-1
Configured evaluation evidence
  • Benchmark: FinGov-Agent-Safety
  • Governance Compliance Rate target: 100%
  • 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

Action Space
  • AutoGen · Required
  • RAI · Optional
Policy Serving
  • Claude 4 Opus · Optional

Evaluation contract

  • hitl_trigger_validation · Weight 1
  • governance_compliance_rate · Weight 1

Metrics

  • governance_compliance_rate · Target 1

Benchmarks

  • FinGov-Agent-Safety
  • security

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON transaction order details

    Output: JSON execution log with approval state

    Execution harness: Python

Evaluation tools

  • AutoGen · Multi-Agent Framework
  • Claude 4 Opus · Reasoning LLM

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Auditable Agentic Financial Execution Guardrail System with AutoGen and Claude 4 Opus