Back to evaluations
Public evaluation
cyber_incident_analysis
Evaluates cyber incident severity classification accuracy and attack vector identification.
Evaluation type
task based
Challenge
Cyber Defense Threat Analyzer using Google ADK and VESSL AI
Difficulty
Advanced
Rigor
Unspecified
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_incident_analysis
Parses log entries to determine MITRE attack vector and severity level.
Input format
JSON containing log_raw_text, system_id, and anomaly_flags
Output format
JSON with threat_category (string), mitre_technique (string), severity (CRITICAL|HIGH|MEDIUM|LOW), and f1_score_flag (boolean)