Manuscript setup

Skill sokolmarek/researcher/skills/manuscript-setup

Create a new academic manuscript project with an organized folder structure. Triggers when user says: 'new manuscript', 'start a paper', 'create manuscript', 'new research project', 'set up paper', 'begin writing paper', 'initialize manuscript', 'scaffold the paper project', 'turn this draft into a manuscript project'. Creates the LaTeX or Word project skeleton with one file per section, a bibliography folder, a figures folder, and a config file, including when an existing draft is about to be moved into it. Always use this skill when the user wants to start a new academic writing project, even without the word manuscript.From its SKILL.md

Install

npx -y skills add sokolmarek/researcher --skill manuscript-setup

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

3 stars3 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

7.6 KB, as published. Nobody here has run it

Manuscript Setup

Creates a structured manuscript project in the working directory.

Workflow

Elicit project details from user (or accept defaults):
- Title (required)
- Author list with affiliations (required)
- Target journal (optional: affects template, word limits, required sections)
- Output format: latex (default), word, or both
- Citation style: apa7 (default), ieee, chicago, vancouver, mla
- Paper type: imrad (default), review, case-study, conference, theoretical
Create folder structure:

manuscript/
├── main.tex                    # Master document with \input{} includes
├── abstract.tex                # Abstract (structured or unstructured)
├── introduction.tex            # Introduction section
├── methods.tex                 # Methods / Materials & Methods
├── results.tex                 # Results section
├── discussion.tex              # Discussion section
├── conclusion.tex              # Conclusion section
├── acknowledgments.tex         # Acknowledgments
├── appendix.tex                # Appendix (optional sections)
├── references/
│   └── library.bib             # Bibliography database
├── figures/                    # Figure files (.pdf, .png, .eps)
│   └── .gitkeep
├── tables/                     # Standalone table files (optional)
│   └── .gitkeep
└── config.yaml                 # Project metadata and settings

If Word mode, create equivalent structure:

manuscript/
├── sections/
│   ├── abstract.md
│   ├── introduction.md
│   ├── methods.md
│   ├── results.md
│   ├── discussion.md
│   └── conclusion.md
├── references/
│   └── library.bib
├── figures/
├── tables/
└── config.yaml

Compile the DOCX with the shipped script, templates/word/build-docx.js. Install its one dependency first (package-lock.json is tracked, so npm ci is the right command); a clean checkout has no node_modules/ and the script will refuse to run without it:

cd templates/word && npm ci && cd -
node templates/word/build-docx.js --manuscript <dir> --out paper.docx

Today this generates a title page, numbered headings, paragraphs, and bullet lists from sections/*.md. Tables, figures, tracked changes, and comments are specified in templates/word/article-imrad.md but not implemented yet, so do not tell the user those are working.

If "both" mode, create both structures side by side.

main.tex Template

The generated main.tex must:

Use appropriate document class based on target journal (default: article)
Include standard packages: amsmath, graphicx, hyperref, booktabs, natbib or biblatex
Set citation style from config
Use \input{} for each section file
Include proper \bibliography{} command
If journal specified: use journal's LaTeX class if available (e.g., elsarticle, IEEEtran)

Read templates/latex/article-imrad.tex for the default template.

config.yaml Format

title: "Paper Title"
authors:
  - name: "First Author"
    affiliation: "University"
    email: "[email protected]"
    corresponding: true
    orcid: "0000-0002-1825-0097"     # optional; validated, never fabricated
    ror: "https://ror.org/042nb2s44" # optional institution ROR ID; validated
    credit:                          # optional CRediT roles; each validated
      - "conceptualization"
      - "writing - original draft"
  - name: "Second Author"
    affiliation: "Institute"
journal:
  name: ""                    # Target journal name
  class: "article"            # LaTeX document class
  word_limit: null             # Word limit if known
citation_style: "apa7"
output_format: "latex"         # latex | word | both
paper_type: "imrad"            # imrad | review | case-study | conference | theoretical
created: "2026-04-09"
status: "drafting"             # drafting | review | revision | final

Optional contributor metadata (ORCID, ROR, CRediT)

Each author entry may carry three optional, machine-checkable fields. All three are validated and are NEVER fabricated: if a value is missing, leave it out; if a supplied value is malformed, it is rejected with an actionable message rather than guessed or "corrected" to a nearby valid one.

orcid: an ORCID iD (16 digits in four groups). The last character is an ISO 7064 mod 11-2 check digit, so a mistyped iD is caught rather than silently accepted. Bare (0000-0002-1825-0097), compact, or full-URL forms are all accepted and normalized to https://orcid.org/....
ror: a Research Organization Registry ID for the affiliation, matched against ROR's published pattern (https://ror.org/0 plus six base32 characters and two check digits) and normalized to the canonical URL. Put it on the author for a single affiliation, or inside each affiliation entry when an author has several.
credit: one or more roles from the CRediT taxonomy's fixed 14 terms (conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, visualization, writing - original draft, writing - review and editing). Matching is case-insensitive and tolerant of hyphen-vs-space; a role outside the taxonomy is rejected.

The kernel validates these (researcher_core.export.validate_orcid, validate_ror, validate_credit_role) and can build a validated Contributor from a config entry (contributor_from_mapping) and emit a JATS <contrib-group> from contributors (to_jats_contrib_group) with <contrib-id contrib-id-type="orcid">, <institution-id institution-id-type="ror">, and CRediT <role> elements, for journals that ingest structured metadata. In LaTeX output, templates/latex/article-imrad.tex renders an author's ORCID next to their name when one is present, and omits it gracefully otherwise.

Section File Initialization

Each section .tex file should be created with:

A \section{} header
A brief comment explaining what goes in this section
Placeholder % TODO: markers for key content areas
Example structure comments based on paper type

Example introduction.tex:

\section{Introduction}
% Structure: Context → Problem → Gap → Contribution → Paper overview

% TODO: Broad context and motivation (1-2 paragraphs)

% TODO: Specific problem statement and existing approaches (1-2 paragraphs)

% TODO: Research gap: what is missing or unsolved (1 paragraph)

% TODO: Your contribution: what this paper does (1 paragraph)
% State contributions as a numbered or bulleted list if appropriate.

% TODO: Paper overview: brief roadmap of remaining sections (1 paragraph)

After Creation

Inform user of created structure
Suggest next steps: "You can now /researcher:draft-section introduction to start writing, or search literature first"
If journal specified, mention any journal-specific requirements detected