Back to evaluations
Draft evaluation
Build an Automated Cyber Incident Triage System with AutoGen and Pydantic AI — evaluation
Evaluates the incident response agent's capacity to accurately map threat indicators to MITRE tactics and output structured playbooks.
Evaluation type
task based
Challenge
Build an Automated Cyber Incident Triage System with AutoGen and Pydantic AI
Difficulty
Advanced
Rigor
Not declared
The author has not specified a rigor level.
Evaluation overview
How the linked challenge is judged: tasks, benchmarks, and criteria count.
Tasks
1
Benchmarks
0
Criteria
0
Task templates
Inputs and expected outputs.
Task 1
cyber_threat_triage_task
Parse raw SIEM alert text and output MITRE technique ID and recommended isolate action
Input format
Unstructured text payload containing firewall and system execution logs
Output format
JSON object with threat_actor_class, mitre_technique_id, severity, and remediation_action