nature-statistics (nature-skills): a Claude skill that audits how a manuscript reports its statistics

nature-statistics makes Claude check how your manuscript reports statistics and rewrite weak methods, results and figure legends without inventing numbers.

Installs: 12.2K; on skills.sh
Stars: 51.7K; on GitHub
Files: 10; besides SKILL.md
Updated: Nov 2026; last repo commit

Give Claude a Statistical analysis section, results text, figure legends or reviewer comments, and it audits what is reported. It separates what was measured, which unit was analysed, and what was claimed. It treats the independent experimental unit as the default n, and flags cells, fields or repeated readings counted as separate samples.

You get a review with P0, P1 and P2 issues, ready-to-paste revised text and a list of facts only you can supply, marked AUTHOR_INPUT_NEEDED. It never invents p values, sample sizes or tests. It does not rerun your analysis unless you supply data and ask. It has extra checks for the flagship Nature journal and for Nature Machine Intelligence.

Good for

Authors preparing a Statistical analysis section for a Nature-family journal
Researchers answering a reviewer who says the statistics are insufficient
Students checking n, error bars and p values in figure legends

Install nature-statistics

Claude Code

Copy this prompt.

Install the nature-statistics skill in Claude Code. Run this command:

npx skills add https://github.com/yuan1z0825/nature-skills --skill nature-statistics --skill nature-shared

Then check that it loaded and tell me how to use it.

2
Paste it into Claude Code. It shows the change and asks before running it.

Requires Node.js 18 or newer (README).

This installs nature-statistics together with nature-shared, because SKILL.md points to files in nature-shared for Nature Machine Intelligence checks and for consistency sweeps. The README says nature-shared is a shared support package and that installing one skill does not add its dependencies. The Claude app ZIP upload is not offered here, since it would need both folders.

Built from the package data we hold. Not yet tested by us.

What's in the skill

When it loads

Audit or improve manuscript statistical reporting, including experimental units, replication, uncertainty, tests, and figure statistics. Use for 统计审查、统计方法小节、图注统计 and reviewer concerns; compute new analyses only when requested with data.

Sections in SKILL.md

Default stance
Accepted inputs
Workflow
Output format
Red lines
Related files
Source hierarchy

Key facts

Author: yuan1z0825 (nature-skills repo owner)
Repo: yuan1z0825/nature-skills (skills/nature-statistics)
Installs on skills.sh: 12,248 as of 2026-11-02
GitHub stars: 51,696 (whole repo) as of 2026-11-02
Last repo commit: 2026-11-01
Folder contents: 11 files including SKILL.md: 6 reference files, 2 READMEs, manifest.yaml and agents/openai.yaml
License: Apache-2.0 (repo license, crawl data)
skills.sh audits: Gen Agent Trust Hub pass, Socket pass, Snyk pass

What it includes

Decides between audit, rewrite, draft, reviewer help, figure alignment or reanalysis.
1. Classify the task
Lists groups, time points, endpoints, blocking, randomization and exclusions.
2. Extract the design
Separates independent units from technical replicates and repeated measures.
3. Define n and replication
Records test, assumptions, correction, effect size and uncertainty per claim.
4. Map claims to analyses
Checks nested data, many comparisons, interactions, regression and outliers.
5. Common failure modes
Checks Methods and Results, with extra rules for Nature and NMI.
6. Reporting completeness
Checks legends, stars, error bars, box and violin plots, and source data.
7. Figure statistics
Writes conservative text without upgrading association to causation.
8. Draft or revise
Runs a reviewer checklist with severity labels and open author questions.
9. Final QA
Bans invented numbers, final test advice without a clear design, and hidden weak results.
Red lines

From the official docs

Questions

What does nature-statistics do?

It audits and rewrites statistical reporting in a manuscript. It checks the design, the unit of analysis, replicates, tests, corrections, effect sizes, uncertainty and figure statistics. It returns a prioritized issue list and conservative text you can paste.

How do I install it?

Use the install box above. It installs this skill with nature-shared, which SKILL.md links to for journal format and consistency files. Installing nature-statistics alone leaves those links without a target.

Will it run my statistics?

Not by default. SKILL.md says it is a reporting and review skill, and it computes new analyses only when you give it data and explicitly ask. It also says it is not a substitute for a statistician.

Does it make up numbers?

No. SKILL.md says it must not invent p values, sample sizes, degrees of freedom, confidence intervals, software versions, correction methods, exclusion rules or power calculations. Missing facts are listed as AUTHOR_INPUT_NEEDED.

Which journals does it cover?

Its general guidance follows Nature Portfolio reporting standards. It adds exact checks for the flagship Nature journal, and for Nature Machine Intelligence through nature-shared files. SKILL.md says to follow a more specific journal instruction if it conflicts.

Does it need an API key?

No key is mentioned. It reads the text, tables or data you give it.

Is it safe?

Everything you paste, including unpublished results, goes to your AI tool, so use it as you would any manuscript help. The skill bundles no scripts. The linked nature-shared package has a consistency-check Python script, which a consistency sweep can run on your manuscript files.

What is the output format?

A review with scope, major issues tagged P0, P1 or P2, a ready-to-paste revision, an AUTHOR_INPUT_NEEDED list and a reviewer-risk note. For a clean drafting request it returns draft text with notes on n, tests, multiple comparisons and software.

Information

updated Nov 2, 2026

Repository: yuan1z0825/nature-skills
Folder: skills/nature-statistics

yuan1z0825/nature-skills skills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 4

nature-skills README · github.com
nature-statistics SKILL.md · github.com
nature-statistics SKILL.md · github.com
nature-statistics on skills.sh · skills.sh