A complete academic research Skill suite covering the full research workflow from paper reading to writing and reviewing. Supports Claude Code, ChatGPT/Codex CLI, and Gemini CLI.
Skill Routing Table
Based on the user's intent, load the corresponding sub-Skill:
| Trigger Condition | Sub-Skill | Path | Description |
|---|---|---|---|
| Read paper, explain paper, paper reading, can't understand this one | Paper Reading | paper-reading/SKILL.md | Great-Grandma character paper walkthrough (Traditional Chinese) |
| Want ideas, brainstorm, research direction, what to do next | Idea Generation | idea-generation/SKILL.md | Three-stage ideation: diverge → search → converge |
| Experiment design, ablation, baseline, what experiments to run | Experiment Design | experiment-design/SKILL.md | Experiment design and planning |
| Mathematical proof, prove, theorem, derivation | Proof Writer | proof-writer/SKILL.md | Theoretical derivation and mathematical proof |
| Write paper, paper writing, improve my paper, LaTeX | Paper Writing | paper-writing/SKILL.md | Paper writing (top conference standard) |
| review, peer review, what would reviewers say, can this get in | Paper Review | paper-review/SKILL.md | 4-step academic peer review |
| Professor analysis, professor fit, choosing an advisor, cold email, application strategy | Professor Fit Analyser | professor-fit-analyser/SKILL.md | Professor fit analysis and application strategy |
Guidelines: When the user's request matches the trigger conditions above, read the SKILL.md at the corresponding path and follow its instructions. If the user's needs span multiple Skills, process them in order according to the Pipeline sequence below.
Skill Pipeline
professor-fit-analyser ─┐
↓
paper-reading ──→ idea-generation ──→ experiment-design
│ │
↓ ↓
paper-review ←── paper-writing ←──── proof-writer
│ ↑
└─────────────────┘ (revision cycle)
Language Conventions
- Default language: Traditional Chinese (analysis, explanation, discussion)
- English scenarios: LaTeX generation, formal review output, mathematical symbols and theorem names
- Academic vocabulary: Refer to shared/chinese-academic-glossary.md to ensure consistency
Shared Resources
- shared/chinese-academic-glossary.md — Chinese-English academic vocabulary reference
- shared/conference-standards.md — Format standards for various top conferences
- shared/researcher-philosophies.md — Researcher philosophy and writing style
Cross-Platform Installation
This suite conforms to the Agent Skills open standard, and can be used on the following platforms:
Claude Code
# 方法一：Clone 到 skills 目錄
git clone <repo-url> ~/.claude/skills/academic-research
# 方法二：在專案中使用
git clone <repo-url> .claude/skills/academic-research
ChatGPT / Codex CLI
git clone <repo-url> ~/.codex/skills/academic-research
# 或在專案中
git clone <repo-url> .codex/skills/academic-research
Gemini CLI
git clone <repo-url> ~/.gemini/skills/academic-research
# 或在專案中
git clone <repo-url> .gemini/skills/academic-research
Universal Method
Simply copy this repository's directory into your AI agent's skills directory. The root directory SKILL.md serves as the entry point, and the agent will automatically discover all sub-skills through the */SKILL.md pattern.