Skip to main content
Back to challenges
Challenge

COSMIC AML Detection Workflow with OpenAI Agents SDK & Braintrust

Financial institutions need automated anti-money laundering sharing pipelines to detect cross-bank illicit networks. Build a COSMIC collaborative sharing intelligence pipeline using OpenAI Agents SDK to analyze financial logs and flag suspicious networks with audited confidence scoring.

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.

Implement an automated anti-money laundering analysis workflow using OpenAI Agents SDK and Braintrust for prompt evaluation.

Delivery guide

How work is evaluated

Evaluation

Evaluates AML detection precision, recall, and structured evaluation logging via Braintrust.

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

valid_json_structure

Ensures risk_score and flagged_entities exist

Binary check

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

Dimension 2

f1_score

F1 Score on anomaly identification • target: 0.9 • 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

  • Build a multi-tool agent using OpenAI Agents SDK

  • Implement COSMIC data parsing and anomaly scoring

  • Evaluate AML detection accuracy using Braintrust

Resources and assets

Reference links and supporting material

Dataset notes

Synthetic bank ledger records mimicking Singapore financial transactions for COSMIC AML evaluation.

Challenge setup

Evaluates AML detection precision, recall, and structured evaluation logging via Braintrust.

Challenge input

JSON containing customer logs and transaction graphs

OpenAI Agents SDK

Agent orchestration

Braintrust

Prompt evaluation and observability

Evaluated output

JSON containing risk_score, flagged_entities, and reasoning

Evaluation checks
  • Ensures risk_score and flagged_entities exist
  • F1 Score on anomaly identification • target: 0.9 • range: 0-1
Configured evaluation evidence
  • Benchmark: FinAML-Bench
  • F1 Score target: 0.9
  • 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
  • RAI · Optional
  • OpenAI · Optional
Observation
  • Braintrust · Required
Reward / Eval
  • Braintrust · Required

Evaluation contract

  • valid_json_structure · Weight 1
  • f1_score · Weight 1

Metrics

  • f1_score · Target 0.9

Benchmarks

  • FinAML-Bench
  • security

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing customer logs and transaction graphs

    Output: JSON containing risk_score, flagged_entities, and reasoning

    Execution harness: Python

Evaluation tools

  • OpenAI Agents SDK · Agent Framework
  • Braintrust · AI Workflow Evaluation

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about COSMIC AML Detection Workflow with OpenAI Agents SDK & Braintrust