santifs/thesis-writing-skill

Thesis Writing Skill is an installable skill for AI coding and writing agents, such as Codex or Claude Code, that helps students and researchers at any level work through the process of writing an empirical thesis or dissertation. It covers the full arc from choosing a topic to final review, and is described as format-neutral: it works equally with Markdown, Word documents, Google Docs prose, or LaTeX, so users are not required to switch tools. The skill addresses several distinct phases of thesis work. In the early stage it helps narrow a broad subject into concrete research questions, hypotheses, a defined scope, and a statement of significance. From there it assists with building chapter and section outlines suited to the conventions of undergraduate, master's, or doctoral empirical research. During drafting it can generate section text from evidence, data tables, figures, and cited sources. It also handles review at two levels: the overall structure and argument flow of the document, and the quality of individual paragraphs using a claim-evidence-interpretation model. One specific discipline it maintains is the distinction between Results and Discussion chapters. Results should report findings factually without interpretation, and Discussion should interpret those findings without restating raw results. The skill is built to preserve that line. The repository contains four files that work together: a main skill definition, a structure guide, a writing guide, and a review checklist. All four should stay in the same folder because the main file references the others by relative path. Installation is either through a command-line installer script for Codex users or by manually copying the folder into the skills directory for whichever agent you use. Usage requires explicitly invoking the skill when making a request to the agent, for example asking the agent to use the thesis-writing skill when reviewing a specific chapter.