Back to evaluations
Public evaluation

ThreatDetectionAndClassification

The evaluation module will assess the LangGraph agent system's ability to accurately detect, classify, and orchestrate remediation for simulated cyber threats. Performance metrics will focus on accuracy, response time, and the quality of generated remediation plans.

Evaluation type
task based
Challenge
Cyberthreat Orchestrator Agent
Difficulty
Advanced
Rigor
Unspecified

Evaluation overview

How the linked challenge is judged: tasks, benchmarks, and criteria count.

Tasks
2
Benchmarks
0
Criteria
0

Task templates

Inputs and expected outputs.

Task 1

ThreatDetectionAndClassification

Evaluate the agent's ability to detect a given threat pattern and classify its severity.

Input format

{"log_entry": "string", "context": "string"}

Output format

{"threat_detected": "boolean", "threat_type": "string", "severity": "string"}

Task 2

RemediationPlanGeneration

Evaluate the agent's ability to generate a coherent and actionable remediation plan for a classified threat.

Input format

{"threat_type": "string", "severity": "string", "affected_systems": ["string"]}

Output format

{"remediation_plan": "list of strings", "estimated_impact": "string"}