Skip to main content
skill

scientific-critical-thinking · davila7 — Claude Code skill: what it does

Checks methodology, statistics and bias in research papers using GRADE and Cochrane ROB. Useful for paper reviews and systematic reviews.

Install into Claude Code · claude-code-templates All methods and platforms ↓
npx claude-code-templates@latest --skill scientific/scientific-critical-thinking
Open on GitHub

What scientific-critical-thinking does

The skill works through a paper in layers. It starts with study design: does it fit the research question, and can it support the causal claims being made. Then it moves to validity — internal, external, construct and statistical — covering randomization, confounder control and blinding.

A separate pass targets bias. It looks for confirmation bias, HARKing, publication bias, selection and survivorship effects, p-hacking and outcome switching. Evidence quality gets scored against GRADE and Cochrane risk of bias. Bundled references include references/scientific_method.md, references/experimental_design.md and references/common_biases.md, plus scripts/generate_schematic.py for diagrams.

The skill can call Read, Write, Edit and Bash, so schematic generation runs inside the session. That part needs Python, since the script ships with the skill. If you only want a text critique of a paper, you can skip the diagram step entirely.

What it helps with

  • assess a paper's methodology before citing it
  • spot confounders and bias in an observational study
  • check whether the design supports causal claims
  • score evidence quality with GRADE
  • audit statistics and outcome switching in a contested publication

Example prompts for the agent

  • Review the methodology of this paper and flag weak points in the design
  • Check this study for selection bias and p-hacking
  • Run a Cochrane risk of bias assessment on this trial

How to install scientific-critical-thinking

1claude-code-templates

npx claude-code-templates@latest --skill scientific/scientific-critical-thinking

2Skills CLI (Vercel)

npx skills add davila7/claude-code-templates --skill scientific-critical-thinking -a claude-code

3Manually: put the files in place

~/.claude/skills/scientific-critical-thinking/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/scientific-critical-thinking/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about scientific-critical-thinking

How do I install scientific-critical-thinking in Claude?

Use the templates command: npx claude-code-templates@latest --skill scientific/scientific-critical-thinking. Alternatively run npx skills add davila7/claude-code-templates --skill scientific-critical-thinking -a claude-code.

What does scientific-critical-thinking do?

It evaluates research papers across methodology, experimental design, statistical validity, bias and confounding. Evidence quality is graded with GRADE and Cochrane ROB.

What do I need for schematic generation?

The scripts/generate_schematic.py script runs through Python and writes images into figures/. It is invoked via the Bash tool.

Similar skills

Everything for Science and research →

scientific-critical-thinking by other authors comes first.

scientific-critical-thinking K-Dense-AI Evaluates scientific claims, research methodology, and evidence quality. Applies GRADE and Cochrane Risk of Bias frameworks. CX Codex +1 2,400installs scientific-brainstorming K-Dense-AI Structures early-stage research ideation through independent idea generation, transparent scoring, and adversarial review. Useful for ranking hypotheses without premature consensus. CX Codex +1 2,300installs scientific-schematics K-Dense-AI Creates PNG drafts of scientific diagrams and runs a separate quality review. Use it for neural network architectures, flowcharts, system diagrams, and biological pathways. CX Codex +1 2,200installs nature-reviewer Yuan1z0825 The skill runs a Nature-style referee assessment of a scientific manuscript: three independent reviewer reports plus a post-review synthesis. Useful before you submit. CX Codex +1 13,900installs literature-search-arxiv google-deepmind Searches arXiv for papers and preprints, returns metadata and abstracts, and downloads full text as PDF, HTML or LaTeX source. CX Codex +1 3,400installs scientific-toolkit-skill zLanqing A skill for research computing: MATLAB/Octave, scientific Python, statistics and publication figures. Useful when you need to compute something, plot a spectrum or check a formula. CX Codex +1 816installs scientific-manuscript-review lyndonkl A skill that walks a scientific manuscript through eight review passes: structure, sections, claim clarity, language and journal formatting. CX Codex +1 479installs

From the same set205

The whole set →

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor