incident-investigation

Investigate HSE incidents and generate de-identified reports with root cause analysis.

1|Updated Jun 14, 2026
One-click install
npx skills add https://github.com/ashley-eyekyam/hse-leadership-skills --skill incident-investigation-ashley-eyekyam
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: incident-investigation
Source: https://github.com/ashley-eyekyam/hse-leadership-skills/tree/main/skills/incident-investigation
Command: npx skills add https://github.com/ashley-eyekyam/hse-leadership-skills --skill incident-investigation-ashley-eyekyam

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

Generic AI-generated safety paperwork is often rejected by regulators: it contains vague task descriptions, controls that default to PPE, actions with no assigned owners, causes that stop at "worker error", and leaked personal or health data. This skill eliminates those failures by forcing specificity before drafting, applying the hierarchy of controls, de-identifying all sensitive data, and running domain-expert review before any output is shared.

Core Features & Use Cases

  • Structured Intake & Specificity Enforcement: Asks one question at a time to gather exact site, activity, evidence, jurisdiction and control details, refusing to proceed on vague or missing inputs.
  • Mandatory De-Identification First: Scrubs all personal/health identifiers (names, IDs, addresses, medical details) to stable role labels, aggregates small injury cells (<5) to protect privacy, and holds the re-identification key separately from all outputs.
  • Validated Root-Cause Analysis (RCA): Supports five A7-validated methods (5-Whys, ICAM, SCAT, Fishbone, Swiss-Cheese) that enforce systemic reach (no stopping at individual blame), with every cause traced to a numbered evidence item.
  • Hierarchy-of-Control CAPA: Generates corrective and preventive actions ranked by the hierarchy of controls, each with a named owner, ISO due date, measurable outcome, and traceability to a specific root or contributing cause.
  • Jurisdiction-Specific Reportability Checks: Provides cited, conservative verdicts on incident reportability for RIDDOR (UK), OSHA 29 CFR 1904 (USA), India state accident forms, and EU frameworks, surfacing the verdict even when an incident is not reportable.
  • Mandatory Quality & SME Review: Runs pre-output validation checks and a dual-lens SME review (lead incident investigator + regulatory specialist) before any output is delivered, marked as decision-support requiring competent-person sign-off. Use cases include investigating injuries, near-misses, property or environmental loss; conducting root cause analysis; building corrective action plans; and determining regulatory reportability for HSE events.

Quick Start

Use the incident-investigation skill to investigate the Line 4 power press injury at Plant 3, run an ICAM root cause analysis, and check reportability for Maharashtra, India.

Frequently Asked Questions about incident-investigation

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate a defensible HSE incident report with proper root cause analysis?

To generate a defensible HSE incident report, the skill enforces structured intake, applies validated root cause analysis methods like 5-Whys or ICAM, and ranks corrective actions by the hierarchy of controls before expert review.

How do I de-identify personal health data in an incident investigation report?

To de-identify personal health data in an incident investigation, the skill scrubs names, IDs, and medical details, replacing them with stable role labels and aggregating small injury cells to protect privacy before analysis.

How do I check if a workplace injury is reportable under OSHA or RIDDOR regulations?

To check if a workplace injury is reportable under OSHA or RIDDOR, the skill runs jurisdiction-specific reportability checks and provides cited, conservative verdicts from its embedded knowledge base for regulatory reporting.

What is the best way to create a CAPA plan that follows the hierarchy of controls?

The best way to create a CAPA plan following the hierarchy of controls is to trace corrective actions to specific root causes, assigning named owners, ISO due dates, and measurable outcomes while ranking controls above PPE.

Can I use 5-Whys and ICAM methods for incident root cause analysis without stopping at worker error?

Yes, you can use 5-Whys and ICAM for incident root cause analysis; the skill enforces systemic reach across five validated methods, ensuring every cause is traced to numbered evidence and preventing analysis from stopping at worker error.

Does this incident investigation process work for near-misses and environmental loss events?

Yes, this incident investigation process works for near-misses and environmental loss events, applying the same de-identification, root cause analysis, and hierarchy-of-controls CAPA generation used for workplace injuries and property damage.