Can it detect p-hacking or selective reporting?

Yes, the skill includes specific checkpoints to identify suspicious clustering of p-values, outcome switching, and lack of pre-registration transparency.

Does it support the GRADE framework for evidence quality?

Yes, it applies GRADE considerations to downgrade or upgrade evidence quality based on risk of bias, consistency, directness, and precision.

How does this skill help with peer review?

It provides a structured framework to assess methodology, identify selection and measurement biases, and evaluate the statistical soundness of a manuscript.

Is it useful for data scientists and ML engineers?

Absolutely, it helps evaluate whether data-driven models are overfitted, if assumptions are met, and if the underlying design supports causal claims.

What frameworks are used for bias detection?

It uses a systematic approach covering cognitive, selection, measurement, and analysis biases to ensure a comprehensive audit of research quality.

Scientific Critical Thinking

Name: Scientific Critical Thinking
Author: Zehong-Wang

byZehong-Wang

0•

データサイエンスとML

Evaluates research rigor and methodology to assess the validity of scientific claims and experimental designs.

Scientific Critical Thinking is a specialized skill designed to enhance Claude's ability to perform deep, systematic evaluations of research papers and scientific claims. It provides a robust framework for assessing experimental design, internal and external validity, statistical power, and potential biases including cognitive, selection, and measurement errors. By leveraging industry-standard tools like GRADE and the Cochrane Risk of Bias (ROB) framework, this skill helps researchers, academics, and data scientists identify confounding variables and logical fallacies, ensuring that conclusions are supported by high-quality, reproducible evidence.

主な機能

01Systematic Bias Detection (Selection, Measurement, and Analysis)

020 GitHub stars

03Comprehensive Methodology Critique (Design, Validity, and Control)

04Statistical Analysis Evaluation (Power, P-values, and Effect Sizes)

05Logical Fallacy Identification in Scientific Arguments

06Evidence Quality Assessment (GRADE & Cochrane ROB frameworks)

ユースケース

01Evaluating the validity of data-driven claims in technical reports

02Conducting systematic literature reviews and meta-analyses

03Reviewing academic manuscripts for peer review and quality control

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add zehong-wang/kosmos scientific-critical-thinking

For use in Claude.ai and ChatGPT

Download Skill