Skip to main content
Back to challenges
Challenge

Cyber Attack Vector Scoring Engine using LlamaIndex and Google Gemini

Sophisticated AI-driven cyber exploits require rapid assessment of corporate threat vectors to underwrite cyber insurance accurately. You will build a multimodal cyber threat scoring engine using LlamaIndex and Google Gemini to ingest breach reports and source code scans. Achieve an unweighted precision score exceeding 0.90 on vulnerability exposure ratings.

Machine LearningHosted 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 enterprise RAG index using LlamaIndex and Google Gemini to calculate cyber liability metrics.

Delivery guide

How work is evaluated

Evaluation

Assess accuracy and retrieval precision of cyber vulnerability exposure ratings.

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

citation_validity

Citations map to actual indexed document names.

Binary check

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

Dimension 2

precision

Precision score of hazard classifications • 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 multimodal vector indices with LlamaIndex and Google Gemini

  • Ingest technical breach notifications and codebase dependency trees

  • Implement structured threat exposure retrieval workflows

  • Quantify dynamic cyber insurance loss potentials

Resources and assets

Reference links and supporting material

Dataset notes

Repository of public vulnerability advisories, synthetic tech stack configurations, and security breach summaries.

Challenge setup

Assess accuracy and retrieval precision of cyber vulnerability exposure ratings.

Challenge input

JSON with company tech stack and query string

LlamaIndex

Indexing, vector retrieval, and structured data querying.

Google Gemini

Multimodal ingestion and deep semantic analysis.

Evaluated output

JSON with exposure rating and key vulnerability citations

Evaluation checks
  • Citations map to actual indexed document names.
  • Precision score of hazard classifications • target: 0.9 • range: 0-1
Configured evaluation evidence
  • Benchmark: CyberRiskBench-2026
  • Precision 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
  • Google · Required
  • Google Gemini · Optional
  • Llama Index · Optional

Evaluation contract

  • citation_validity · Weight 1
  • precision · Weight 1

Metrics

  • precision · Target 0.9

Benchmarks

  • CyberRiskBench-2026
  • security

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON with company tech stack and query string

    Output: JSON with exposure rating and key vulnerability citations

    Execution harness: Python

Evaluation tools

  • LlamaIndex · RAG Framework
  • Google Gemini · LLM Provider

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Cyber Attack Vector Scoring Engine using LlamaIndex and Google Gemini