Back to evaluations
Public evaluation
osha_safety_audit
Evaluates precision and recall of OSHA non-compliance safety violation detection.
Evaluation type
task based
Challenge
CrewAI OSHA Plant Safety Audit Crew with Specialized Expert Roles
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
osha_safety_audit
Evaluates CrewAI safety report against ground truth safety violations in work order text.
Input format
JSON object with work_orders array and safety_log_text
Output format
JSON object with violations_found, osha_clause_citations, and overall_risk_score