How does this skill handle multiple simultaneous changes?

It utilizes interaction analysis to identify exponential effects where two or more changes combine to create a failure that neither would cause individually.

Does this skill account for external factors?

Yes, the methodology includes 'Natural changes' like traffic growth, seasonal patterns, and upstream dependency updates that occur outside of direct deployments.

What is the primary goal of the Change Analysis skill?

The goal is to identify the root cause of a problem by isolating what specifically changed between a system's last known working state and its current failing state.

When should I use Change Analysis over other RCA methods?

Use Change Analysis when a system was working fine until a specific point in time. It is less effective for systems that never worked correctly or for purely random failures.

Change Analysis Root Cause Investigation

Name: Change Analysis Root Cause Investigation
Author: lev-os

bylev-os

0•

Analytics & Monitoring

Identifies the root cause of system failures by systematically comparing the current failed state to a previous problem-free baseline.

Change Analysis is a specialized diagnostic skill designed to help developers and SREs pinpoint the exact origin of system regressions. By moving beyond simple correlation, this skill provides a structured framework for mapping all modifications—including code deployments, environmental shifts, and external dependency updates—that occurred between a working baseline and the onset of an incident. It is particularly effective for troubleshooting post-deployment bugs, performance degradation, and complex system behaviors where the primary question is 'what changed?'

Key Features

01Temporal comparison between 'working' and 'failing' system states

02Comprehensive change mapping (Mandated, Natural, and Caused changes)

03Structured hypothesis testing through rollback and reproduction

040 GitHub stars

05Interaction analysis for non-additive simultaneous changes

06Causal relationship classification and evidence-based validation

Use Cases

01Analyzing quality shifts in system outputs where the behavior differs from established baselines

02Diagnosing performance degradation or latency spikes that appeared after a specific deployment

03Investigating sudden increases in error rates or service failures in production environments

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add lev-os/agents change-analysis

For use in Claude.ai and ChatGPT

Download Skill