a5c-ai Root Cause Analysis Skill
Overview
The Root Cause Analysis (RCA) skill provides structured methodologies and tools for conducting thorough root cause investigations for incidents and problems. This skill supports multiple analysis frameworks including 5-Whys, Fishbone/Ishikawa diagrams, and Fault Tree Analysis to identify underlying causes and drive permanent corrective actions.
Capabilities
5-Whys Analysis
- Facilitate structured 5-Whys questioning
- Document causal chain progression
- Identify stopping criteria for root causes
- Generate 5-Whys analysis reports
- Track branching cause paths
- Validate root cause identification
Fishbone/Ishikawa Diagrams
- Generate fishbone diagram structures
- Organize causes by category (6M model)
- Support custom category definitions
- Create visual diagram outputs (Mermaid, SVG)
- Facilitate brainstorming sessions
- Document contributing factors
Fault Tree Analysis
- Build fault tree logical structures
- Define AND/OR gate relationships
- Calculate probability propagation
- Identify minimal cut sets
- Generate fault tree visualizations
- Support quantitative risk analysis
Contributing Factor Identification
- Analyze multiple cause categories
- Identify systemic vs. proximate causes
- Document environmental factors
- Assess human factors contributions
- Evaluate process deficiencies
- Identify technology failures
Root Cause Confidence Scoring
- Calculate confidence levels for identified causes
- Assess evidence strength for each factor
- Rate cause probability and impact
- Generate confidence intervals
- Prioritize causes by confidence
- Document uncertainty and assumptions
RCA Documentation Generation
- Generate comprehensive RCA reports
- Create executive summaries
- Document timeline of events
- Include evidence and data references
- Produce corrective action recommendations
- Generate lessons learned documentation
Corrective Action Tracking
- Track corrective action implementation
- Monitor action completion status
- Verify effectiveness of fixes
- Generate action status reports
- Link actions to root causes
- Track action ownership and deadlines
Usage
Conduct 5-Whys Analysis
skill: rca-analysis
action: five-whys
parameters:
incident_id: "INC-2025-001"
problem_statement: "Customer dashboard failed to load for 2 hours"
analysis:
why_1:
question: "Why did the dashboard fail to load?"
answer: "The API gateway returned 503 errors"
why_2:
question: "Why did the API gateway return 503?"
answer: "Backend services were unresponsive"
why_3:
question: "Why were backend services unresponsive?"
answer: "Database connection pool was exhausted"
why_4:
question: "Why was the connection pool exhausted?"
answer: "A slow query was holding connections for extended periods"
why_5:
question: "Why was there a slow query?"
answer: "Missing database index on newly added column"
root_cause: "Missing database index caused slow queries that exhausted connection pool"
Generate Fishbone Diagram
skill: rca-analysis
action: fishbone-diagram
parameters:
problem: "High customer churn rate in Q4"
categories:
people:
- "Insufficient CSM staffing"
- "High CSM turnover"
- "Lack of proactive outreach training"
process:
- "No early warning system"
- "Delayed QBR scheduling"
- "Reactive-only engagement model"
technology:
- "Health score not capturing usage decline"
- "CRM data quality issues"
- "No automated alerting"
measurement:
- "Churn identified too late"
- "NPS not correlated to churn"
environment:
- "Economic downturn pressure"
- "Competitor aggressive pricing"
output_format: mermaid
Perform Fault Tree Analysis
skill: rca-analysis
action: fault-tree
parameters:
top_event: "Complete service outage"
tree:
gate: OR
events:
- name: "Infrastructure failure"
gate: AND
events:
- name: "Primary DC failure"
probability: 0.001
- name: "DR failover failure"
probability: 0.01
- name: "Application failure"
gate: OR
events:
- name: "Code deployment error"
probability: 0.02
- name: "Configuration error"
probability: 0.015
calculate:
- top_event_probability
- minimal_cut_sets
- critical_events
Generate RCA Report
skill: rca-analysis
action: generate-report
parameters:
incident_id: "INC-2025-001"
report_type: comprehensive
include:
- executive_summary
- incident_timeline
- impact_assessment
- analysis_methodology
- root_causes
- contributing_factors
- corrective_actions
- lessons_learned
format: markdown
Track Corrective Actions
skill: rca-analysis
action: track-actions
parameters:
rca_id: "RCA-2025-001"
actions:
- id: "CA-001"
description: "Add missing database index"
owner: "database-team"
due_date: "2025-01-20"
status: completed
verification: "Query performance improved from 30s to 50ms"
- id: "CA-002"
description: "Implement connection pool monitoring"
owner: "sre-team"
due_date: "2025-01-25"
status: in_progress
Process Integration
This skill integrates with the following customer experience processes:
| Process | Integration Points |
|---|---|
| problem-management.js | Core RCA methodology, known error documentation, permanent fixes |
| itil-incident-management.js | Post-incident review, PIR documentation, preventive actions |
| closed-loop-feedback.js | Customer complaint root causes, systemic issue identification |
Analysis Frameworks
6M Categories (Manufacturing)
- Manpower (People)
- Methods (Process)
- Machines (Technology)
- Materials
- Measurement
- Mother Nature (Environment)
6S Categories (Service)
- Surroundings
- Suppliers
- Systems
- Skills
- Safety
- Standards
Dependencies
- RCA templates and frameworks
- Analysis methodology guidelines
- Incident/problem data access
- Corrective action tracking system
- Visualization libraries for diagrams
- Historical RCA database for patterns
Best Practices
- Start Quickly: Begin RCA while evidence is fresh and available
- Blame-Free Culture: Focus on systemic issues, not individuals
- Multiple Perspectives: Include diverse stakeholders in analysis
- Evidence-Based: Support conclusions with data and logs
- Actionable Outputs: Ensure corrective actions are specific and measurable
- Follow Through: Track actions to completion and verify effectiveness
- Learn and Share: Document lessons learned for organizational knowledge
Shared Potential
This skill is a strong candidate for extraction to a shared library with applicability to:
- DevOps/SRE (incident post-mortems)
- QA Testing (defect analysis)
- Software Architecture (failure analysis)
- Manufacturing Operations
- Healthcare Quality