Skip to main content
Back to challenges
Challenge

Design Security Audit Agent Teams with OpenAI Agents SDK and E2B

Create a robust government repository security auditing system using the official OpenAI Agents SDK. This multi-agent system uses Claude Opus 4.6.6 for high-reasoning tasks and vulnerability assessment. The workflow leverages E2B for secure code execution environments to run static analysis tools, while Vercel v0 generates frontend status dashboards. Interaction with the security team is augmented by Coplay AI to handle conversational reporting.

Special Purpose AgentsHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Orchestrate specialized security agents using OpenAI Agents SDK to identify and report repository vulnerabilities.

Delivery guide

How work is evaluated

Evaluation

Validation of vulnerability detection accuracy within the sandboxed environment.

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

SecurityCheck

Ensure no false negatives in critical vulnerabilities

Binary check

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

Dimension 2

DetectionRate

Ratio of caught vulnerabilities • 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 the OpenAI Agents SDK pattern for defining multi-turn autonomous agents

  • Implement E2B sandboxes for isolating vulnerability testing environments

  • Design Claude Opus 4.6.6 reasoning chains for complex security audits

  • Generate dynamic security compliance dashboards using Vercel v0 integration

  • Implement natural language feedback loops using Coplay AI interfaces

Resources and assets

Reference links and supporting material

Dataset notes

Sample data for 1 tasks

Challenge setup

Validation of vulnerability detection accuracy within the sandboxed environment.

Challenge input

GitHub Repository URL

E2B

Cloud sandboxes for AI agents.

OpenAI

OpenAI AI model provider

Vercel v0

AI app and interface builder.

Evaluated output

Vulnerability Report

Evaluation checks
  • Ensure no false negatives in critical vulnerabilities
  • Ratio of caught vulnerabilities • target: 1 • range: 0-1
Configured evaluation evidence
  • DetectionRate target: 100%
  • 1 public reference case
Configured runtime
  • Docker execution harness

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

Environment
  • E2B · Required
Action Space
  • E2B · Required
  • OpenAI · Optional
  • Vercel v0 · Optional

Evaluation contract

  • SecurityCheck · Weight 1
  • DetectionRate · Weight 1

Metrics

  • DetectionRate · Target 1

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: GitHub Repository URL

    Output: Vulnerability Report

    Execution harness: Docker

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Design Security Audit Agent Teams with OpenAI Agents SDK and E2B