Skip to main content
Loading navigation
Back to challenges

Cyber Catastrophe Bond Trigger & Portfolio Accumulation Monitor with Mastra AI

Managing accumulation risk in Cyber Catastrophe Bonds requires real-time aggregation of widespread cloud outage events across global insured portfolios. You will build an autonomous cyber catastrophe monitoring system using Mastra AI in TypeScript to track systemic cloud downtime signals against bond trigger thresholds. The system must process telemetry with sub-1 second latency and zero unparsed events.

Machine LearningHosted by Vera
Sign in free to start

What you are building

Build a real-time cyber catastrophe bond accumulation and trigger monitoring pipeline in TypeScript using Mastra AI workflows and agent tools.

How work is evaluated

Evaluation

Evaluates Mastra AI cyber cat bond monitoring workflow on cloud downtime telemetry streams.

Datasets

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

How submissions are scored

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

loss_positive

Estimated loss must be positive float.

Binary check

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

Dimension 2

workflow_execution_success

Percentage of events processed cleanly without workflow failure. • 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

  • Define Mastra AI workflows (`@mastra/core/workflows`) with structured steps for cyber incident telemetry.

  • Build custom Mastra tools that query cloud availability metrics and cloud outage logs.

  • Implement real-time aggregation models for estimating industry economic loss from hyperscaler downtime.

  • Output automated alerts for Cyber Cat Bond triggers when estimated claims breach attachment thresholds.

Resources

Dataset notes

Synthetic stream of 150 cloud outage incidents detailing downtime durations, affected cloud services, and global user impacts.

Challenge setup

Evaluates Mastra AI cyber cat bond monitoring workflow on cloud downtime telemetry streams.

Challenge input

JSON containing event_id, cloud_provider, downtime_hours, affected_regions, portfolio_market_share

Mastra AI

Provides structured, type-safe workflow engines for serverless TypeScript applications.

Evaluated output

JSON containing total_estimated_industry_loss_usd, cat_bond_impacted (BOOLEAN), alert_level

Evaluation checks
  • Estimated loss must be positive float.
  • Percentage of events processed cleanly without workflow failure. • target: 1 • range: 0-1
Configured evaluation evidence
  • Benchmark: MastraCyberBench-2026
  • Workflow Execution Success target: 1
  • 1 public reference case
Configured runtime
  • JavaScript execution harness
  • javascript sandbox (unavailable on Versalist)

Shows the first task, up to three tools, and up to three evidence items.

View technical recipe

Configured tools

Action Space
  • Mastra AI · Required
  • RunPod · Optional
Policy Serving
  • ERNIE 4.0 · Optional
Orchestration
  • Mastra AI · Required

Evaluation contract

  • loss_positive · Weight 1
  • workflow_execution_success · Weight 1

Metrics

  • workflow_execution_success · Target 1

Benchmarks

  • MastraCyberBench-2026
  • security

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing event_id, cloud_provider, downtime_hours, affected_regions, portfolio_market_share

    Output: JSON containing total_estimated_industry_loss_usd, cat_bond_impacted (BOOLEAN), alert_level

    Execution harness: JavaScript

Evaluation tools

  • Mastra AI · TypeScript Agent & Workflow Framework

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Cyber Catastrophe Bond Trigger & Portfolio Accumulation Monitor with Mastra AI