skills.rest
Mzane0803
Mzane0803COMMUNITY
Explore all skills by this author →
Website𝕏@MZane0803GitHub

scientific-writing

Draft and audit scientific manuscripts with evidence provenance tracking and local consistency checks.

Updated Aug 12, 2026
Search Tags:#manuscript#scientific writing#reporting guidelines#authorship#research integrity#citation audit#evidence provenance
One-click install
npx skills add https://github.com/Mzane0803/latent-minds-skills-marketplace --skill scientific-writing-mzane0803
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: scientific-writing
Source: https://github.com/Mzane0803/latent-minds-skills-marketplace/tree/main/plugins/research-writing/skills/scientific-writing
Command: npx skills add https://github.com/Mzane0803/latent-minds-skills-marketplace --skill scientific-writing-mzane0803

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve? Scientific manuscripts risk fabricated citations, inconsistent numbers, undisclosed AI use, and confidentiality breaches when drafted with generic AI assistance. This Skill enforces a fail-closed workflow where every factual claim maps to human-verified evidence before submission. ## Core Features & Use Cases - Evidence-bound drafting: Assign claim IDs and evidence IDs to every factual or numeric statement, with local CLI audits that reject unverified or untagged content. - Reporting-guideline routing: Select applicable guidelines (CONSORT 2025, PRISMA 2020, STROBE, TRIPOD+AI, and others) by study design and record non-scoring coverage. - Consistency and integrity checks: Validate numeric facts, units, denominators, methods-results mappings, reference identifiers, authorship criteria, CRediT roles, and AI-use disclosures with offline Python tools. - Use Case: A research team drafting a randomized trial manuscript scaffolds a workspace, registers verified sources, drafts with claim markers, then runs the linter and claim audit to confirm no placeholders, unverified citations, or confidentiality gate violations remain before submission. ## Quick Start Ask the AI to scaffold a new manuscript workspace for your study design and draft a section using only verified evidence from your source manifest.

Frequently Asked Questions about scientific-writing

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

FAQPage Schema
How do I draft a scientific manuscript without fabricated citations?▼

Register each source in a source manifest with an evidence ID, then draft using claim and evidence markers like [claim:C001] [evidence:E001]. The audit_claims.py script rejects any claim lacking verified evidence, and sources are only marked verified after a human opens and confirms them.

How to check consistency between methods and results sections?▼

Record numeric facts, methods, outcomes, and results in a consistency manifest JSON file, then run check_consistency.py. It flags unit mismatches, cross-section value conflicts, numerator-denominator errors, and results lacking declared methods.

What reporting guideline should I use for my study design?▼

Run select_reporting_guidelines.py with your study design, such as randomized_trial or systematic_review. It routes to current guidelines like CONSORT 2025, PRISMA 2020, STROBE, or TRIPOD+AI, but the selection is non-scoring and you must consult the official checklist.

Does the reference checker verify DOIs and PMIDs online?▼

No, check_references.py validates only local syntax and duplicates for DOI, PMID, PMCID, ISBN, and URL identifiers without network resolution. A human must still compare every identifier against the opened source.

Can I upload my unpublished manuscript to external AI services?▼

Not without explicit authorization and a documented policy review covering journal, institutional, funder, consent, and legal requirements. The bundled tools run entirely offline with no network calls, so local processing is the default safe boundary.

What are the limitations of the manuscript linter?▼

The linter flags placeholders, overstatement, causal language in observational designs, and possible sensitive content by line number, but sensitive-content warnings are review prompts, not a de-identification certificate. Only accountable humans can set submission_ready to true.

Related Skills

Explore More Education & Research Skills→
brainbrewlabsbrainbrewlabs
7

source-gathering

Collect and organize credible research sources with APA citations.

Official
Intermediate
pradyumnasagarpradyumnasagar
6

response-to-reviewers

Constructs structured revision responses to reviewer comments for journal submissions.

Community
Intermediate
TerryFYLTerryFYL
8

paper-pipeline

Orchestrate a nine-skill pipeline from raw data to submission package.

Community
Advanced
MustDemirMustDemir

thesis-writer

Write academic thesis paragraphs with APA7 citation verification and proof protocols.

Community
Advanced