Supervisor-Skills
Twelve skills for research work: idea evaluation, literature review, paper text, reviewer rebuttals and figures
- 12,379 ★ on the repository
- Last commit
Extended access
Why this level
- The skills call scholarly search, web search and the public Crossref, Semantic Scholar and arXiv APIs
- The agent reads your drafts and experiment data and writes files with text and figures
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/supervisor-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The skills call scholarly search, web search and the public Crossref, Semantic Scholar and arXiv APIs
- The agent reads your drafts and experiment data and writes files with text and figures
This is third-party code. Review the repository files before installing.
What it does
Supervisor-Skills is a set of twelve skills built by a lecturer at the Hong Kong University of Science and Technology (Guangzhou) from publishing experience at SIGMOD, VLDB, ICML and NeurIPS. The skills follow the stages of working on a paper. `idea-evaluator` scores an idea on five dimensions, `deep-research` builds a literature review with every citation verified, `tech-paper-template` and `benchmark-paper-template` lay out the paper's logic, `intro-drafter` writes the introduction and `paper-writer` drafts the sections. Next come `paper-polish` for language editing, `pre-submission-reviewer` for a draft check through a reviewer's eyes and `rebuttal-guidance` for replies to comments. `figure-designer` advises on drawing a motivating example, a solution overview and a results chart, while `drawio-reconstruction` rebuilds a figure from an image into a `.drawio` file. A handbook with theory and paper breakdowns sits alongside.
Who it is for. PhD students and computer science researchers writing a conference paper and working in Claude Code, Cursor or Codex.
Good fit when
- A paper idea has appeared and should be scored against criteria before months go into experiments
- A literature review is needed before the introduction, with every citation checked against the source
- A draft is ready and gets a conference reviewer's check before submission
- Reviews have arrived and each comment needs a reply plan with tone and evidence
- A figure from another paper or slide should become an editable Draw.io file
Not a fit when
- The paper is outside computer science: the author's experience and templates are built on that field
- You want the agent to invent experiment results: the skills insist on your own materials
- The project is commercial: the CC BY-NC-SA license forbids commercial use
Example request
Evaluate my paper idea with the five-dimension scheme, then build a literature review on the topic with every citation verifiedLimitations
The CC BY-NC-SA 4.0 license permits only non-commercial use and requires derivatives to be shared on the same terms. The `paper-writer` and `deep-research` skills find sources through the agent's scholarly search, web search or public APIs such as Crossref, Semantic Scholar and arXiv; without them a skill works from your materials with no new citations. The project calls itself early-stage, and the description and news in the main README are in Chinese, with the English version in `README.en.md`. The author suggests an order for the skills, but you assemble them into a chain yourself.
How to disable. Delete the folders you installed from the agent's skills directory (`.claude/skills/`, `.codex/skills/` or `.cursor/skills/`). If you cloned the repository, you can delete it entirely.
FAQ
How do I install Supervisor-Skills in Claude Code?
Send the agent the line from the README: Help me install Supervisor-Skills from https://github.com/HKUSTDial/Supervisor-Skills with Skills. The agent places the skills in the directory and you pick the ones you need.
Which skills are included?
Twelve: `idea-evaluator`, `vibe-research-workflow`, `deep-research`, `tech-paper-template`, `benchmark-paper-template`, `intro-drafter`, `paper-writer`, `paper-polish`, `pre-submission-reviewer`, `figure-designer`, `rebuttal-guidance` and `drawio-reconstruction`.
Does `paper-writer` invent facts and citations?
No, the skill writes only what your materials, retrieved sources or common field knowledge without numbers or names support. An unverified citation is removed from the text.
How do I redraw a figure from a paper in Draw.io?
Give the agent the image and call `drawio-reconstruction`: the skill builds a `.drawio` file from native elements, simple SVG icons and cropped images, then shows a preview to compare with the original. It edits only the files you name.
Can the skills be used in commercial work?
No, the CC BY-NC-SA 4.0 license allows only non-commercial use. A commercial task requires the authors' permission.
Discussion
No questions or reviews yet.
Related
SkillsSkill with 30 prompt templates for research papers in English and Chinese, from literature review to grant proposals
Eight skills for finished papers and theses in LaTeX and Typst: compile, references, language polish, cover letters and a reviewer-style audit
A skill set for research work in the style of Nature journals: finding and reading papers, writing, statistics, figures, reviewer responses
Ten research skills for Claude Code and Codex: figures, conference and journal papers, grants, peer review and code release
A skill for researchers who draft in Chinese: drafts and results become an English manuscript, with citations and claims checked
Three skills for academic writing: a paper schema for image generation, rewriting text to sound less machine-made, and drafting from scratch
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| Supervisor-Skills (this card) | no | 12,379 | |
| Academic Writing Skills | no | 549 | a set of academic writing skills without idea evaluation or reviewer rebuttals |
| Deep Research Skills | no | 2,538 | a single deep-research skill covering one stage of paper work |
README in short
The main README is in Chinese: it explains why the author built the project and lists news, including the skills being deployed on Baidu and Doubao platforms. It then gives the repository structure with the handbook and the `skills` folder, a table of twelve skills, a one-line quick start for an agent and links to the related DataMagic and DeepEye projects.
README badge
Original SKILL.md
--- name: paper-writer description: >- Drafts publishable paper prose from the author's own materials, from a single paragraph to a full manuscript, across STEM and non-STEM fields. Every factual claim traces to user input, verified retrieval, or field common knowledge; citations pass an independent verification ladder; delivery is clean prose with zero placeholder tags. Use when the user asks to write a section, turn an idea or results into paper text, or draft a full paper. license: CC-BY-NC-SA-4.0 --- # Paper Writer ## Overview This skill turns a researcher's ideas and materials into submittable paper prose. One paragraph works; a full manuscript works. It is not an outline generator and not a writing tutor. It writes the text, under one non-negotiable condition: every written word must be traceable to evidence. The author owns the substance (ideas, designs, results, conclusions); this skill owns the linguistic realization of that substance, and nothing more. ## Hard rules These hold for every task this skill performs: 1. Every factual claim has one of three origins: the user's materials, this session's verified retrieval, or field common knowledge that carries no numbers, names, or comparisons. Model memory is never a source. 2. Delivered prose contains zero bracketed placeholder tags of any kind. A claim without a source is resolved by searching, rewriting, or deleting, never by tagging. 3. Concrete details the user did not provide are not written: no invented scenarios, mechanisms, magnitudes, procedures, or identifiers. 4. Claim strength never exceeds evidence strength. 5. Real results and planned or expected results are phrased differently and never mixed. 6. Internal planning never leaks into the output. The full discipline, including the evidence hierarchy and the Evidence Map: see references/evidence-discipline.md.