Supervisor-Skills

Supervisor-Skills is an open-source collection of research Skills from Yuyu Luo, assistant professor at the Hong Kong University of Science and Technology (Guangzhou), and his lab (GitHub organization HKUSTDial). The README says that from his own PhD to supervising students, he kept seeing graduate students stuck on the last mile of research, so the team wrote up its experience publishing and reviewing at venues such as SIGMOD, VLDB and NeurIPS as a Chinese handbook, then distilled it into 12 Skills: evaluating ideas, in-depth literature research, drafting introductions, evidence-gated paper writing, meaning-preserving polishing, thinking templates for technical and benchmark papers, a pre-submission check from a reviewer's point of view, figure design advice, rebuilding reference images as editable Draw.io files, and rebuttal guidance. What sets it apart is that each Skill has a matching handbook chapter, so you can learn the method first and then have the AI follow it.

Who it's for:graduate students starting out in research and researchers writing or revising papers, especially in data science and AI.

Before you start: copy the Skills you need from the skills folder into your AI tool, or ask your agent to install them as the README describes; the README says Claude, DeepSeek, Kimi and others all work. The Draw.io rebuild runs Python scripts on your computer and needs Pillow and the Draw.io desktop app, and it asks before installing anything. Your papers and images are processed by the AI model you use. The license is CC BY-NC-SA 4.0: non-commercial use only, and adaptations must be shared under the same license.

Based on the project's public documentation. Not tested by AI Inspiration World.