crisai-research-methodology

Guide hypothesis testing, root cause analysis, and experimental flag validation in crisAI.

Updated Apr 18, 2026
One-click install
npx skills add https://github.com/crissdiamond/crisAI --skill crisai-research-methodology
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: crisai-research-methodology
Source: https://github.com/crissdiamond/crisAI/tree/main/.claude/skills/crisai-research-methodology
Command: npx skills add https://github.com/crissdiamond/crisAI --skill crisai-research-methodology

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This Skill provides a structured methodology for turning research hunches into accepted results within the crisAI platform, ensuring thorough investigation and validation.

Core Features & Use Cases

  • Root Cause Analysis: Offers a systematic approach to identifying and validating root causes of issues.
  • Hypothesis Testing: Encourages precise prediction of outcomes before running tests, with a focus on trace events and gate outcomes.
  • Idea Lifecycle Management: Guides the development of ideas through various stages, including live failure sessions, defect decomposition, and ADR creation.
  • Experiment Flag Management: Provides guidelines for introducing new behaviors with opt-in switches and validation against live sessions.
  • Validation and Retirement: Ensures experiments are validated across all execution modes and ends appropriately with flips or documented retirements.

Quick Start

Load the crisai-research-methodology skill to start or review an investigation, propose a fix, or plan a feature flag.

Frequently Asked Questions about crisai-research-methodology

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

FAQPage Schema
What is the best way to run root cause analysis for an experimental flag?

Root cause analysis for an experimental flag requires a systematic approach to identify and validate underlying issues. This methodology guides you through defect decomposition and live failure sessions to ensure thorough investigation before proposing a fix.

How do I test a hypothesis before rolling out an experiment flag?

Hypothesis testing requires predicting outcomes before running tests, focusing on trace events and gate outcomes. You can structure predictions and validate experimental flags against live sessions to ensure reliable results.

How do I manage the lifecycle of an idea from hunch to accepted result?

Managing an idea lifecycle involves stages like live failure sessions, defect decomposition, and ADR creation. This process transforms a hunch into an accepted result by enforcing structured workflow modes and evidence bar standards.

When do I need to validate and retire an experiment flag?

Validation and retirement of an experiment flag are needed after testing across all execution modes. Experiments should end with flips or documented retirements to ensure new behaviors introduced via opt-in switches are properly concluded.

Can I use this research methodology without prior knowledge of crisAI workflows?

Using this research methodology requires adherence to structured workflow modes and evidence bar standards within crisAI. Familiarity with hypothesis testing, trace events, and ADR creation is necessary to effectively validate experimental flags.