Use to get PNAS Nexus references right — note that PNAS Nexus accepts references in any readable style at submission (unlike flagship PNAS's strict numbered-by-appearance), so the job is internal consistency, completeness, resolvable in-text citations, and the [dataset] tag for data/software. Prepare a single consistent (preferably numbered) style for production; confirm the final style on acceptance.
References are inconsistent, incomplete, or mix several styles.
In-text citations don't all resolve to the list (or vice versa).
Datasets/software are cited as bare URLs rather than formal references.
You are about to submit and want the reference list clean.
The PNAS Nexus rule that surprises people: format-neutral at submission
PNAS Nexus states, verbatim, "You may format references in any readable style at submission." This is a deliberate contrast with flagship PNAS, which enforces a numbered, in-order-of-appearance style up front. So at PNAS Nexus, do not burn effort forcing a specific house style for initial submission — instead make the references clean, complete, consistent, and resolvable.
The final published reference style is 待核实 — OUP/PNAS-portfolio journals typically render a numbered, in-order-of-appearance list at production, but the official page only guarantees submission-stage neutrality. Prepare a single consistent numbered style (it converts cleanly to most production styles) and confirm the exact format on acceptance.
What actually matters at submission (the real checklist)
One consistent style throughout — pick a reference-manager style (a numbered/Vancouver or PNAS-style works well) and apply it uniformly; do not leave a mix.
Every in-text citation resolves to an entry in the list, and every list entry is cited — no orphans, gaps, or duplicates.
Complete metadata — authors, title, journal/venue, year, volume, pages or article number, and a DOI where one exists.
Author lists complete and accurate (avoid silent "et al." truncation in the list unless the chosen style requires it; reference managers often over-truncate — check).
Data and software cited as references with the [dataset] tag where the journal asks for it (see pnasnexus-data), not buried as URLs in the text.
Preprints clearly labeled as preprints with their repository/DOI.
A safe, convertible default (numbered, in order of appearance)
If you want one style that is clean now and converts easily at production, use numbered, in order of appearance:
In text: numbers in parentheses or brackets — (1), (2, 3), (4–6) — in the order references first appear.
A single numbered reference list in appearance order.
Reference formats (shape — confirm production requirements on acceptance):
Journal article:1. A. B. Author, C. D. Author, Article title. J. Abbrev. Volume, page–page (Year). DOI.
Book:2. A. B. Author, Book Title (Publisher, ed. X, Year).
Book chapter:3. A. B. Author, Chapter title in Book Title, C. D. Editor, Ed. (Publisher, Year), pp. xx–yy.
Dataset:[dataset] 4. A. B. Author, Dataset title. Repository. Accession/DOI. Deposited DD Month Year.
Preprint:5. A. B. Author, Title. Repository [Preprint] (Year). DOI (accessed DD Month Year).
Tooling
Use Zotero/EndNote with a numbered or PNAS-style CSL; do a final manual pass on completeness and author-list truncation (managers often truncate).
Verify every in-text citation resolves and the list has no gaps/duplicates; renumber after any reordering of the text.
Run a DOI-completeness pass — add DOIs wherever they exist.
Output format
【Submission rule】 references format-neutral at submission (PNAS Nexus) — make them clean, not house-styled
【Consistency】 one style applied uniformly? yes/no
【Resolvable】 all in-text ↔ list, no orphans/gaps/duplicates? yes/no
【Completeness】 authors / title / venue / year / vol / pages / DOI present? yes/no
【Data/software refs】 cited with [dataset] tag where required? yes/no
【Production style】 numbered-by-appearance prepared (convertible)? confirm final style on acceptance (待核实)
【Next】 pnasnexus-submission
Anti-patterns
Do not waste pre-submission effort hand-forcing a strict house style — PNAS Nexus accepts any readable style at submission.
Do not ship a list that mixes several styles inconsistently.
Do not leave in-text citations that don't resolve, or list entries never cited.
Do not cite datasets/software as bare URLs — use formal references with the [dataset] tag.
Do not assume the submission style is the final published style — confirm production requirements on acceptance.
1---2name: pnasnexus-citation3description: Use to get PNAS Nexus references right — note that PNAS Nexus accepts references in any readable style at submission (unlike flagship PNAS's strict numbered-by-appearance), so the job is internal consistency, completeness, resolvable in-text citations, and the [dataset] tag for data/software. Prepare a single consistent (preferably numbered) style for production; confirm the final style on acceptance.4---567# Reference Style (pnasnexus-citation)89## When to trigger1011- References are inconsistent, incomplete, or mix several styles.12- In-text citations don't all resolve to the list (or vice versa).13- Datasets/software are cited as bare URLs rather than formal references.14- You are about to submit and want the reference list clean.1516## The PNAS Nexus rule that surprises people: format-neutral at submission1718PNAS Nexus states, verbatim, *"You may format references in any readable style at submission."* This is a **deliberate contrast with flagship PNAS**, which enforces a numbered, in-order-of-appearance style up front. So at PNAS Nexus, **do not burn effort forcing a specific house style for initial submission** — instead make the references **clean, complete, consistent, and resolvable**.1920> The **final published** reference style is **待核实** — OUP/PNAS-portfolio journals typically render a **numbered, in-order-of-appearance** list at production, but the official page only guarantees submission-stage neutrality. Prepare a single consistent numbered style (it converts cleanly to most production styles) and confirm the exact format on acceptance.2122## What actually matters at submission (the real checklist)2324- [ ] **One consistent style** throughout — pick a reference-manager style (a numbered/Vancouver or PNAS-style works well) and apply it uniformly; do not leave a mix.25- [ ] **Every in-text citation resolves** to an entry in the list, and every list entry is cited — no orphans, gaps, or duplicates.26- [ ] **Complete metadata** — authors, title, journal/venue, year, volume, pages or article number, and a **DOI** where one exists.27- [ ] **Author lists** complete and accurate (avoid silent "et al." truncation in the list unless the chosen style requires it; reference managers often over-truncate — check).28- [ ] **Data and software cited as references** with the **`[dataset]`** tag where the journal asks for it (see `pnasnexus-data`), not buried as URLs in the text.29- [ ] **Preprints** clearly labeled as preprints with their repository/DOI.3031## A safe, convertible default (numbered, in order of appearance)3233If you want one style that is clean now and converts easily at production, use **numbered, in order of appearance**:3435- In text: numbers in parentheses or brackets — **(1)**, **(2, 3)**, **(4–6)** — in the order references first appear.36- A **single numbered reference list** in appearance order.37- Each entry: authors, article title, abbreviated journal/venue, volume, pages/article number, year, DOI.3839Reference formats (shape — confirm production requirements on acceptance):4041- **Journal article:** `1. A. B. Author, C. D. Author, Article title. J. Abbrev. Volume, page–page (Year). DOI.`42- **Book:** `2. A. B. Author, Book Title (Publisher, ed. X, Year).`43- **Book chapter:** `3. A. B. Author, Chapter title in Book Title, C. D. Editor, Ed. (Publisher, Year), pp. xx–yy.`44- **Dataset:** `[dataset] 4. A. B. Author, Dataset title. Repository. Accession/DOI. Deposited DD Month Year.`45- **Preprint:** `5. A. B. Author, Title. Repository [Preprint] (Year). DOI (accessed DD Month Year).`4647## Tooling4849- Use Zotero/EndNote with a numbered or PNAS-style CSL; do a final manual pass on completeness and author-list truncation (managers often truncate).50- Verify every in-text citation resolves and the list has no gaps/duplicates; renumber after any reordering of the text.51- Run a DOI-completeness pass — add DOIs wherever they exist.5253## Output format5455```56【Submission rule】 references format-neutral at submission (PNAS Nexus) — make them clean, not house-styled57【Consistency】 one style applied uniformly? yes/no58【Resolvable】 all in-text ↔ list, no orphans/gaps/duplicates? yes/no59【Completeness】 authors / title / venue / year / vol / pages / DOI present? yes/no60【Data/software refs】 cited with [dataset] tag where required? yes/no61【Production style】 numbered-by-appearance prepared (convertible)? confirm final style on acceptance (待核实)62【Next】 pnasnexus-submission63```6465## Anti-patterns6667- **Do not** waste pre-submission effort hand-forcing a strict house style — PNAS Nexus accepts any readable style at submission.68- **Do not** ship a list that mixes several styles inconsistently.69- **Do not** leave in-text citations that don't resolve, or list entries never cited.70- **Do not** cite datasets/software as bare URLs — use formal references with the `[dataset]` tag.71- **Do not** assume the submission style is the final published style — confirm production requirements on acceptance.7273---7475**Source:** [`brycewang-stanford/Awesome-Journal-Skills`](https://github.com/brycewang-stanford/Awesome-Journal-Skills) → `PNAS-Nexus-Skills/skills/pnasnexus-citation/SKILL.md`
Run npx skillmds add thedixitjain/pnasnexus-citation in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.
Use to get PNAS Nexus references right — note that PNAS Nexus accepts references in any readable style at submission (unlike flagship PNAS's strict numbered-by-appearance), so the job is internal consistency, completeness, resolvable in-text citations, and the [dataset] tag for data/software. Prepare a single consistent (preferably numbered) style for production; confirm the final style on acceptance. It is listed under AI & ML on SkillMD.
This skill has not completed SkillMD's automated safety review yet. Capability flags: makes network calls. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.
This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.
Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.
thedixitjain (@thedixitjain) published this skill. Their other Agent Skills are listed on their SkillMD profile.