CrewAI OSHA Plant Safety Audit Crew with Specialized Expert Roles
Scaling Industry 4.0 applications requires strict adherence to workplace safety standards like OSHA 1910. Construct a collaborative CrewAI team featuring an OSHA Inspector, Safety Engineer, and Operations Lead to audit plant floor inspection logs. Success is measured by achieving 95%+ precision in identifying OSHA non-compliance events across factory floor operational logs.
What you are building
The core problem, expected build, and operating context for this challenge.
Orchestrate a CrewAI multi-agent team with distinct safety inspection roles to audit industrial telemetry and work order logs against OSHA standards.
How work is evaluated
Evaluates precision and recall of OSHA non-compliance safety violation detection.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
citation_present
Ensures output includes accurate 29 CFR 1910 citation strings.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
audit_f1_score
F1 score of identified safety violations against true risk incidents. • target: 0.95 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Formulate specialized CrewAI Agents with explicit roles, goals, and backstories.
Assign structured Tasks for auditing Lockout/Tagout (LOTO) and machine guarding logs.
Synthesize crew findings into an executive plant safety risk score report.
Validate audit findings against published OSHA 1910 standard guidelines.
Reference links and supporting material
Factory maintenance work order logs and incident reports containing embedded safety policy violations.
Challenge setup
Evaluates precision and recall of OSHA non-compliance safety violation detection.
Challenge input
JSON object with work_orders array and safety_log_text
CrewAI
Required framework for building role-playing agent crews with collaborative tasks.
Evaluated output
JSON object with violations_found, osha_clause_citations, and overall_risk_score
- Ensures output includes accurate 29 CFR 1910 citation strings.
- F1 score of identified safety violations against true risk incidents. • target: 0.95 • range: 0-1
- Benchmark: OSHAAuditBench-2026
- Audit F1 Score target: 0.95
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- CrewAI · Required
- crewAI · Optional
Evaluation contract
- citation_present · Weight 1
- audit_f1_score · Weight 1
Metrics
- audit_f1_score · Target 0.95
Benchmarks
- OSHAAuditBench-2026
- security
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON object with work_orders array and safety_log_text
Output: JSON object with violations_found, osha_clause_citations, and overall_risk_score
Execution harness: Python
Evaluation tools
- CrewAI · Agent Framework
Recipe freeze marker
No recipe freeze time is recorded.