Skip to main content
Back to challenges
Challenge

MAS FEAT AI Governance Compliance & Audit Trace Synthesizer with Mastra AI

Construct an automated AI governance and audit compliance agent using Mastra AI to evaluate enterprise model deployment outputs. Audit AI decisions against MAS FEAT principles (Fairness, Ethics, Accountability, Transparency) and generate audit records.

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.

Deploy Mastra AI workflows to systematically test, audit, and log model decisions against Monetary Authority of Singapore FEAT AI governance principles.

Delivery guide

How work is evaluated

Evaluation

Evaluates detection of fairness bias and generation of FEAT audit reports.

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

bias_correctly_flagged

Ensures model bias below 0.80 ratio is flagged as non-compliant

Binary check

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

Dimension 2

ratio_accuracy

Accuracy of mathematical disparate impact calculation • 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

  • Master TypeScript agent workflow orchestration using Mastra AI framework primitives

  • Implement FEAT alignment evaluators measuring bias metrics across demographic attributes

  • Build dynamic memory and RAG stores in Mastra AI to track long-term model audit logs

  • Export standardized regulator-ready compliance artifacts with cryptographic hash verification

Resources and assets

Reference links and supporting material

Dataset notes

Synthetic batch prediction dataset across 500 simulated loan decisions with group attributes.

Challenge setup

Evaluates detection of fairness bias and generation of FEAT audit reports.

Challenge input

JSON object containing model_prediction_batch with protected_attributes

Mastra AI

Provides typed workflows, vector memory, and observability for TypeScript.

Evaluated output

JSON object with feat_compliant (boolean), disparate_impact_ratio, and violation_summary

Evaluation checks
  • Ensures model bias below 0.80 ratio is flagged as non-compliant
  • Accuracy of mathematical disparate impact calculation • target: 1 • range: 0-1
Configured evaluation evidence
  • Benchmark: AI Governance & Fairness Benchmark
  • Ratio Accuracy target: 1
  • 1 public reference case
Configured runtime
  • JavaScript execution harness
  • javascript 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
  • Mastra AI · Required
  • aiXplain · Optional
  • Banani · Optional
Orchestration
  • Mastra AI · Required

Evaluation contract

  • bias_correctly_flagged · Weight 1
  • ratio_accuracy · Weight 1

Metrics

  • ratio_accuracy · Target 1

Benchmarks

  • AI Governance & Fairness Benchmark
  • general

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON object containing model_prediction_batch with protected_attributes

    Output: JSON object with feat_compliant (boolean), disparate_impact_ratio, and violation_summary

    Execution harness: JavaScript

Evaluation tools

  • Mastra AI · TypeScript Agent Framework

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about MAS FEAT AI Governance Compliance & Audit Trace Synthesizer with Mastra AI