Skip to main content
Loading navigation
Back to challenges

Network Failure Root-Cause Agent with OpenAI Agents SDK and E2B Sandbox

Unannounced OT network switch failures degrade operator situational awareness and trigger cascading control logic trips. Construct an autonomous diagnostic agent using the OpenAI Agents SDK inside an E2B sandbox to parse network log files and pinpoint switch failures. Achieve less than 5 seconds resolution time for cascading incident root causes.

EngineeringHosted by Vera
Sign in free to start

What you are building

Develop an OpenAI Agents SDK diagnostic workflow executed inside E2B sandboxes to isolate network switch outages impacting OT safety loops.

How work is evaluated

Evaluation

Evaluates root-cause accuracy and execution speed inside an isolated E2B sandbox.

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

root_cause_accuracy

Verifies that the identified failed_device_ip matches the true broken switch.

Binary check

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

Dimension 2

diagnosis_latency_sec

Time spent by the agent performing root-cause isolation in E2B. • target: 5 • range: 0-30

Binary check

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

Learning goals

  • Deploy custom tool handlers in OpenAI Agents SDK to parse Syslog and PCAP files

  • Execute safe code isolation and log diagnostics within an E2B sandbox container

  • Map communication loss events to safety instrumented system (SIS) trip conditions

Resources

Dataset notes

Network packet captures (PCAP) and industrial syslog logs representing network switch failure cascading to PLC trips.

Challenge setup

Evaluates root-cause accuracy and execution speed inside an isolated E2B sandbox.

Challenge input

ZIP archive containing network PCAP logs and PLC alarm records

OpenAI Agents SDK

Agent orchestration framework

E2B

Isolated sandboxed code execution

Evaluated output

JSON report specifying failed_device_ip, failure_timestamp, and impacted_loops

Evaluation checks
  • Verifies that the identified failed_device_ip matches the true broken switch.
  • Time spent by the agent performing root-cause isolation in E2B. • target: 5 • range: 0-30
Configured evaluation evidence
  • Benchmark: OT Network Fault Diagnosis
  • Diagnosis Latency Sec target: 5
  • 1 public reference case
Configured runtime
  • Python execution harness
  • Python sandbox (unavailable on Versalist)

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

View technical recipe

Configured tools

Environment
  • E2B · Required
Action Space
  • E2B · Required
  • OpenAI · Optional
  • Agent Protocol · Optional

Evaluation contract

  • root_cause_accuracy · Weight 1
  • diagnosis_latency_sec · Weight 1

Metrics

  • diagnosis_latency_sec · Target 5

Benchmarks

  • OT Network Fault Diagnosis
  • security

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: ZIP archive containing network PCAP logs and PLC alarm records

    Output: JSON report specifying failed_device_ip, failure_timestamp, and impacted_loops

    Execution harness: Python

Evaluation tools

  • OpenAI Agents SDK · framework
  • E2B · agent-system

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Network Failure Root-Cause Agent with OpenAI Agents SDK and E2B Sandbox