hypothesis

Generate and test falsifiable hypotheses for CAPPY Phase 4 investigations.

Updated Aug 23, 2026

Search Tags:#investigation #evidence #root-cause #coherence #hypothesis-generation #falsification #causation-analysis

Ready to run

npx skills add https://github.com/theLightArchitect/cappy-toolkit --skill hypothesis-thelightarchitect

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: hypothesis
Source: https://github.com/theLightArchitect/cappy-toolkit/tree/main/skills/hypothesis
Command: npx skills add https://github.com/theLightArchitect/cappy-toolkit --skill hypothesis-thelightarchitect

SYSTEM DOCUMENTATION & REQUIREMENTS

## What problem does it solve? Phase 4 hypothesis generation and causation analysis for CAPPY investigations—providing a structured, evidence-driven method to identify root causes and validate environment fit against Cortex docs and TAC precedents.

## Core Features & Use Cases

Load investigation state and extract Phase 2-3 data from inv_context.json
Generate 2-4 candidate hypotheses, each falsifiable with explicit evidence
Build and evaluate a Tree of Thoughts, perform adversarial falsification, and perform environment-fit checks
Survive best hypothesis, classify issue type, and plan Phase 5 targets

### Quick Start Load an investigation and run Phase 4 hypothesis generation and falsification to identify the strongest survivor.

Frequently Asked Questions about hypothesis

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

FAQPage Schema

How do I generate falsifiable hypotheses for root-cause analysis?▼

Hypothesis generation for root-cause analysis involves creating 2-4 candidate hypotheses with explicit evidence. You then perform adversarial falsification and environment-fit checks to identify the strongest surviving cause. This structured method ensures your investigation yields actionable targets.

What is a Tree of Thoughts approach for causation analysis?▼

A Tree of Thoughts for causation analysis is a structured evaluation method that branches out potential root causes. It allows you to systematically test and falsify various paths against supporting evidence, ensuring the surviving hypothesis accurately reflects the investigation environment.

How do I test investigation evidence against environment-fit constraints?▼

To test investigation evidence against environment-fit constraints, you apply your Phase 2-3 data to validate root causes. This involves checking candidate hypotheses against environment documentation to ensure the identified causation aligns with system precedents and actual configurations.

What is the best way to plan downstream targets after identifying a root cause?▼

The best way to plan downstream targets after root-cause identification is to classify the issue type and establish actionable objectives. By surviving the best hypothesis from your falsification tests, you directly guide Phase 5 planning and gate decisions for the investigation.

Can I use inv_context.json data for Phase 4 hypothesis generation?▼

Yes, you can use inv_context.json data for Phase 4 hypothesis generation. The Skill specifically loads investigation state and extracts Phase 2-3 data from this file to build falsifiable hypothesis sets and perform environment-fit assessments.

Why does my root-cause analysis fail to produce actionable investigation targets?▼

Root-cause analysis often fails to produce actionable targets when hypotheses are not falsifiable. Without applying a Tree of Thoughts evaluation and adversarial falsification, you lack the evidence-based validation needed to confidently classify issues and plan next steps.