This project addresses the "last-mile problem" in academic research, offering AI-powered assistance from idea generation to paper submission. Targeting early-career researchers, it distills a decade of senior academic experience into structured, executable AI skills, serving as a 24/7 AI research co-pilot.

How It Works

Supervisor-Skills uses a dual-track architecture: theoretical "Guide" materials and practical, executable AI skills. Core research experience from top-tier conferences is "distilled" into structured prompts for LLMs (e.g., Claude, GPT-4). This translates implicit expert knowledge into actionable AI tools, moving beyond generic prompts for specialized research support.

Quick Start & Requirements

Installation involves providing a specific prompt to an AI assistant (e.g., Claude Code, Cursor, Codex): "Help me install Supervisor-Skills from https://github.com/HKUSTDial/Supervisor-Skills with Skills." Key requirements include access to compatible AI assistants. No specific Python version or hardware is detailed.

Highlighted Details

• Structured AI Skills: Executable LLM prompts for idea evaluation, intro drafting, paper templating (technical/benchmark), pre-submission review, and figure design.
• Dual-Track Architecture: Integrates theoretical "Guide" with practical "Skills" for comprehensive research support.
• Expert Knowledge Distillation: Condenses ten years of experience from top AI/Data Science conferences (SIGMOD, VLDB, ICML, NeurIPS).
• Top Conference Case Studies: Detailed analyses of paper writing for ICML, ICLR, and VLDB.

Maintenance & Community

The project is in its "early stages" and actively seeks community contributions (issues, PRs). Contact is via email. Future development includes rebuttal writing and collaboration strategies.

Licensing & Compatibility

Licensed under CC BY-NC-SA 4.0, permitting non-commercial sharing and adaptation with attribution. Commercial use is restricted.

Limitations & Caveats

The project is in its early stages with ongoing development (TODOs). The CC BY-NC-SA 4.0 license restricts commercial use. Effectiveness depends on LLM capabilities and user prompting.