What it does

Turning a thesis into articles is not a reformatting exercise. A thesis demonstrates mastery of a field; an article makes one bounded contribution to a specific scholarly conversation. This skill turns that conversion into an explicit, auditable procedure: it reads the full thesis, decides how many standalone articles it really contains, drafts them one at a time and reviews each draft against Q1 peer-review criteria before handing it over.

Its main feature is calibrated honesty. It will not promise three articles when the thesis supports one, and it does not inflate scores to be encouraging. A thesis with a single dataset and a single research question yields a single article; splitting it artificially produces two thin papers that reviewers reject.

Who it is for

• PhD students moving from a monograph thesis to a thesis by publication.
• Early-career researchers extracting their first articles from a completed dissertation.
• Supervisors who want a structured, critical first pass.

Designed for the social sciences: sociology, political science, economics, management, organisation studies and adjacent quantitative and mixed-methods fields.

How it works

1. Diagnosis. Each candidate article goes through a four-question viability test: does it have its own research question, a distinct theoretical contribution, enough empirical material, and can it be read on its own? Failing two or more means it is not viable. The skill announces how many articles there are (1, 2 or 3) and why, before writing anything.
2. Drafting. Each article is written in full in standard Q1 structure. Only material from the thesis is used, with no invented data, results or citations, but it is rewritten: the thesis scaffolding is removed and the argument is rebuilt around the article’s own gap.
3. Peer review. Each draft is scored 1–5 on seven dimensions (contribution, gap, design, empirical evidence, coherence, writing and independence). It gets a verdict (ready to submit, needs revision or not viable as a Q1 article), the three weaknesses a reviewer would flag with their fixes, and two or three target journals.

How to install it

In Claude (web and desktop): download the repository as a ZIP (Code → Download ZIP) and upload it under Settings → Capabilities → Skills. Requires a paid plan.

In Claude Code or Cowork:

git clone https://github.com/juliancardenasx/thesis-to-articles-skill.git ~/.claude/skills/thesis-to-articles

How to use it

Attach the thesis (PDF or DOCX) and ask, in English or Spanish:

Convert my thesis into publishable articles.

¿Cuántos artículos puedo sacar de mi tesis?

Expect to be told “two articles, not three” more often than not. That is the intended behaviour, not a failure.

Limitations

• Social sciences only: the rubric, split patterns and journal lists follow their conventions.
• Journal lists change every year: check current rankings before submitting.
• It drafts; it does not author. Whoever uses it must check everything against their own data and results, and remains responsible for it.
• Check each journal’s policy on AI use and disclose it where required.

MIT licence. If you use it in published work, cite it using the repository’s CITATION.cff file.

Practice

To try it with a ready-made skill, download Thesis-to-Articles from its repository and upload it following the steps in the video.

More videos on our YouTube channel.

What it is about

The chapter is written in Spanish; original title: “Las élites tecnológicas: la inserción en marcos normativos y redes corporativas”.

The chapter studies the dual embeddedness of tech elites in the age of artificial intelligence: their institutional embeddedness in the frameworks that regulate AI, and their relational embeddedness in corporate networks.

It compares the regulatory models of the United States and the European Union and analyses the interlocking directorates network of the 21 largest technology companies. The result is a tension: regulation is split into two paradigms (the European one, based on risk of use, and the US one, based on computing capacity), yet the networks are integrated through transatlantic ties. Meta, Microsoft, Amazon and Alphabet connect with European tech firms such as Accenture, Adyen and SAP. This divergence between fragmented institutions and integrated networks calls into question whether national frameworks can constrain elites that operate transnationally.

How to cite

Cárdenas, J. and Domínguez-Rubio, Z. (2026). Las élites tecnológicas: la inserción en marcos normativos y redes corporativas. In M. Serna, E. T. dos Reis, I. Gastal Grill and M. Gené (Eds.), Elites políticas e domínios do poder na América Latina (ch. 12). Fino Traço Editora. ISBN 978-85-8054-795-5.

What it is about

An editorial in the Revista Española de Sociología that asks whether we should accept AI reviewing the articles submitted to a journal. It defines and characterises generative AI, reviews tools that already assist research and sets out four possible scenarios: democratisation of research, underuse of AI, rising inequality and digital laggards.

The central argument: AI greatly expands what a researcher can do, but it cannot “hit the streets”. It cannot replace fieldwork or the interpretation of social context. The text closes with concrete proposals: training in AI tools, research-based education and an open debate on the use of AI in managing scientific journals.

How to cite

Cárdenas, J. (2023). Artificial intelligence, research and peer-review: future scenarios and action strategies. Revista Española de Sociología, 32(4), a184. https://doi.org/10.22325/fes/res.2023.184