Supervisor-Skills
A set of AI skills distilling a professor's ten years of research and reviewing experience into structured research assistance, covering a paper's full lifecycle from idea to submission.
Supervisor-Skills packages a decade of publishing and reviewing experience at top data science and AI venues (SIGMOD, VLDB, ICML, NeurIPS) into structured AI skills that a large language model such as Claude or GPT can execute precisely. The project's stated aim is to turn a senior researcher's tacit knowledge and judgment into a tool anyone can use, addressing gaps graduate students commonly face between theory and practice, scarce mentoring time, and uncertainty before submitting a paper.
Skills cover the paper lifecycle end to end, including paper-writer for evidence-gated drafting of the body text, paper-polish for faithful language polishing, deep-research for review-level literature surveys, an upgraded intro-drafter, evaluation and review skills that route by discipline-specific paradigms, and a drawio-reconstruction skill that rebuilds a reference figure into an editable Draw.io diagram.
An accompanying Handbook lays out a systematic guide, from evaluating paper quality from a reviewer's perspective through idea development, the paper-writing process, and figure design, intended to be read alongside using the skills.