research-principles

Document hypotheses, pre-register predictions, and log decisions for auditable experiments.

13|Updated Jan 27, 2026

Search Tags:#documentation #research #audit #principles #hypothesis #rigor #experiments

Ready to run

npx skills add https://github.com/Butanium/claude-lab --skill research-principles

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: research-principles
Source: https://github.com/Butanium/claude-lab/tree/main/plugins/clab/skills/research-principles
Command: npx skills add https://github.com/Butanium/claude-lab --skill research-principles

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Core research principles for hypothesis-driven investigation help teams plan, execute, and review experiments with rigor, transparency, and auditable reasoning.

Core Features & Use Cases

Hypothesis-Driven Exploration: State hypotheses before experiments and pre-register predictions.
Documentation Standards: Log commands, parameters, timestamps, outputs verbatim.
Red-team Your Results: Actively seek disconfirming evidence and test sensitivity of results.
Escalation Ladder: Use staged prompting and validation approaches, from zero-shot to supervised fine-tuning, to ensure reliability.
Reproducibility & Audit: Maintain traceable experiments and decisions across roles (orchestrator, scientists, colleagues, judge).

Quick Start

Document your current hypothesis, pre-register predictions, and log decisions to begin applying these principles.

Frequently Asked Questions about research-principles

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

FAQPage Schema

What is hypothesis-driven research and how does pre-registering predictions work?▼

Hypothesis-driven research requires stating hypotheses before experiments and pre-registering predictions. This ensures investigation rigor and transparency by explicitly documenting expected outcomes prior to execution, preventing post-hoc rationalization.

How do I document experiments to ensure auditability and reproducibility?▼

To ensure auditability, document experiments by logging commands, parameters, timestamps, and outputs verbatim. This rigorous documentation maintains traceable decisions across roles, enabling full reproducibility during research reviews.

What is the best way to red-team research results and test sensitivity?▼

The best way to red-team research results is to actively seek disconfirming evidence and test the sensitivity of your findings. This rigorous approach identifies vulnerabilities in your hypothesis before formal peer review.

How do I use an escalation ladder for validation in research experiments?▼

You use an escalation ladder for validation by applying staged prompting approaches, from zero-shot to supervised fine-tuning. This staged process ensures the reliability of your hypothesis-driven research results.

Can I apply these research principles to orchestrate experiments across different roles?▼

Yes, you can apply these research principles to orchestrate experiments across different roles. The framework supports orchestrators, scientists, and colleagues, maintaining traceable decisions and audits across all research sessions.

When should I not use a staged validation approach for experiments?▼

You should not use a staged validation approach when your experiment lacks a clearly defined, pre-registered hypothesis. Without explicit documentation of predictions, rigorous logging of decisions cannot be effectively validated.