anomaly-investigation

Identify root causes of data anomalies by testing candidate causes.

150|19|Updated Jun 15, 2026

Search Tags:#root cause analysis #investigation log #data anomaly #data-driven diagnosis #feedback signal

Ready to run

npx skills add https://github.com/gaasher/Agent-Loop-Skills --skill anomaly-investigation

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: anomaly-investigation
Source: https://github.com/gaasher/Agent-Loop-Skills/tree/main/loops/anomaly-investigation
Command: npx skills add https://github.com/gaasher/Agent-Loop-Skills --skill anomaly-investigation

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

When you have a specific anomaly in your data, like a metric spike or drop, and need to identify its root cause with certainty, this Skill helps narrow down and confirm the cause based on your data.

Core Features & Use Cases

Root Cause Analysis: Analyze anomalies by systematically testing and eliminating potential causes.
Data-Driven Elimination: Formulate candidate causes and test each against the data to eliminate the ones refuted by the data.
Positive Confirmation: Confirms the root cause by ensuring it passes a positive test that the data supports.
Use Case: If you notice a sudden spike in user sessions, this Skill can help you diagnose whether it's due to a marketing campaign, a bug, or a seasonal event.

Quick Start

Use the anomaly-investigation skill to diagnose the cause of a spike in user sessions.

Frequently Asked Questions about anomaly-investigation

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

FAQPage Schema

How do I find the root cause of a data anomaly like a sudden metric spike?▼

To find the root cause of a data anomaly, you systematically test candidate causes against your dataset. This approach formulates potential drivers, eliminates those refuted by the data, and positively confirms the surviving cause.

What is data-driven diagnosis for investigating a known dataset anomaly?▼

Data-driven diagnosis for a dataset anomaly involves iteratively testing potential causes against your data. It eliminates unsupported candidates and positively confirms the single root cause supported by the data.

How do I systematically test candidate causes to confirm a root cause analysis?▼

You confirm a root cause analysis by iteratively testing each candidate cause against your dataset. This process eliminates causes refuted by the data and ensures the surviving cause passes a positive confirmation test.

Does this root cause analysis approach require Python to run investigation scripts?▼

Yes, executing the analysis scripts for root cause analysis requires Python 3.9 or higher. The scripts test candidate causes against your dataset to eliminate refuted hypotheses and confirm the root cause.

What is the best way to diagnose whether a user session spike is from a bug or campaign?▼

The best way to diagnose a user session spike is to formulate candidate causes like a bug or campaign, test each against your data to eliminate refuted options, and positively confirm the surviving root cause.

When should I not use an iterative elimination approach for anomaly investigation?▼

You should not use iterative elimination for anomaly investigation when you need proactive monitoring rather than reactive diagnosis. This approach is strictly for confirming the root cause of a specific, known data anomaly.