What problem does it solve?
Generic AI-generated safety paperwork is often rejected by regulators: it uses vague tasks, collapses controls to PPE, leaves actions without owners, attributes incidents to generic "worker error", and leaks personal data into widely shared documents. This skill eliminates those failures by enforcing structured intake to capture exact site, activity, and event details, strict API RP 754 compliance for process safety metrics, fail-loud rate discipline to avoid fabricated figures, and mandatory de-identification of all sensitive data before drafting any output.
Core Features & Use Cases
- API RP 754-Aligned KPI Computation: Accurately counts Tier-1 and Tier-2 Process Safety Events (PSEs) from user-provided event facts (never invented) and calculates normalized rates with strict fail-loud denominator discipline (missing or zero work hours results in a [GAP] rate marker, no fabricated figures).
- Leading & Lagging Indicator Framing: Frames Tier-3 (safety system challenges) and Tier-4 (operating discipline) leading indicators alongside lagging PSE counts, with clear distinction from occupational TRIR/LTIFR/DART injury rate metrics.
- Mandatory Compliance Guardrails: Runs structured multi-step intake to enforce input specificity, mandatory de-identification of all personal/health data before drafting, and domain-expert SME review before any output is presented to ensure regulator-defensible results.
- Use Case: A process safety manager at an O&G facility needs to generate a Q2 process safety KPI report for leadership, including Tier-1 PSE counts, normalized rates, and leading indicator performance, with all personal data removed and all figures traceable to source evidence.
Quick Start
Use the process-safety-kpi skill to compute the Tier-1 and Tier-2 process safety event rates for your refinery's Q2 reporting period, providing the total work hours denominator to avoid a [GAP] rate marker.
