Back to evaluations
Public evaluation

physical_tamper_detection

Evaluates physical process attack detection accuracy and structured Pydantic model response validity.

Evaluation type
task based
Challenge
Pydantic AI Physical Process Anomaly Sentinel for Water Utilities
Difficulty
Intermediate
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

physical_tamper_detection

Evaluate incoming water treatment plant sensor telemetry for physical constraint violations.

Input format

JSON with flow_rate_gpm, chlorine_ppm, tank_level_ft, pump_status

Output format

Pydantic output model with physics_violated (bool), anomaly_type (string), threat_level (string)