orange4664/research-skills

research skills

Python

41

14 commits

updated Mar 31, 2026

See the code

README

🧪 Research Skills

A collection of modular AI agent skills for automating academic research workflows — from paper discovery to reproduction.

Each skill is a self-contained tool that can be used:

  • 🖥️ Standalone — as a CLI tool
  • 🤖 As an AI Skill — integrated into Claude Code or other AI coding agents

📦 Available Skills

SkillDescriptionStatus
paper-finderSearch for papers & source code across arXiv, Semantic Scholar, HuggingFace, GitHub✅ Ready
paper-downloaderDownload paper PDFs and clone source code repos✅ Ready
paper-parserParse PDF papers into structured JSON (via MinerU)✅ Ready
code-analyzerDeep AST analysis, training loop dissection, reproducibility scoring✅ Ready
formula2codeConvert LaTeX formulas to PyTorch/NumPy code (15 ML patterns + SymPy pipeline)✅ Ready
code-writerGenerate project scaffolding from paper descriptions (no source code needed)✅ Ready
paper-presenterGenerate Beamer presentations summarizing papers✅ Ready
beamer-skillAcademic Beamer LaTeX presentation lifecycle tool✅ Ready (bundled)
code-reproducerExecute reproduction on remote GPU servers via mcp-ssh✅ Ready
result-analyzerCompare reproduced results against paper figures/metrics. Reports in MD/JSON/Beamer✅ Ready

🚀 Quick Start

git clone https://github.com/orange4664/research-skills.git
cd research-skills

# Use paper-finder
cd paper-finder
pip install -r requirements.txt
python search_paper.py "Attention Is All You Need"

🏗️ Pipeline Architecture

┌──────────────┐    ┌──────────────────┐    ┌──────────────┐
│ paper-finder │───▶│ paper-downloader │───▶│ paper-parser │
│ (search)     │    │ (download)       │    │ (PDF → JSON) │
└──────────────┘    └──────────────────┘    └──────┬───────┘
                                                   │
                    ┌──────────────────┐    ┌───────▼───────┐    ┌──────────────┐
                    │  code-analyzer   │◀───│paper-presenter│───▶│ beamer-skill │
                    │ (AST/ML scoring) │    │ (summarize)   │    │ (LaTeX PPT)  │
                    └────────┬─────────┘    └───────────────┘    └──────────────┘
                             │
          ┌──────────────────┼──────────────────┐
          │                  │                  │
 ┌────────▼─────────┐ ┌──────▼───────┐ ┌────────▼─────────┐
 │  formula2code    │ │ code-writer  │ │ code-reproducer  │
 │ (LaTeX → code)   │ │ (scaffold)   │ │ (train via SSH)  │
 └────────┬─────────┘ └──────┬───────┘ └────────┬─────────┘
          │                  │                  │
          └──────────────────┼──────────────────┘
                    ┌────────▼─────────┐
                    │ result-analyzer  │
                    │ (compare)        │
                    └──────────────────┘

No Source Code? Two new paths:

  • formula2code: Paper equations → PyTorch/NumPy functions
  • code-writer: Paper description → complete project scaffolding (with reference code discovery)

📄 License

MIT License — see LICENSE

🙏 Third-Party

ComponentAuthorLicenseLink
beamer-skillNoi1rMITGitHub
latex2sympy2HuggingFaceMITGitHub
sympytorchpatrick-kidgerApache-2.0GitHub
Reproducibility scoringPapers With CodeML Code Completeness
AST analysis approachPyCG (ICSE'21)Apache-2.0Paper
LaTeX paper writinglatex-paper-skillsMITGitHub

Contributors

orange4664

14 commits

orange4664/research-skills

research skills

Python

41

14 commits

updated Mar 31, 2026

See the code

README

🧪 Research Skills

A collection of modular AI agent skills for automating academic research workflows — from paper discovery to reproduction.

Each skill is a self-contained tool that can be used:

  • 🖥️ Standalone — as a CLI tool
  • 🤖 As an AI Skill — integrated into Claude Code or other AI coding agents

📦 Available Skills

SkillDescriptionStatus
paper-finderSearch for papers & source code across arXiv, Semantic Scholar, HuggingFace, GitHub✅ Ready
paper-downloaderDownload paper PDFs and clone source code repos✅ Ready
paper-parserParse PDF papers into structured JSON (via MinerU)✅ Ready
code-analyzerDeep AST analysis, training loop dissection, reproducibility scoring✅ Ready
formula2codeConvert LaTeX formulas to PyTorch/NumPy code (15 ML patterns + SymPy pipeline)✅ Ready
code-writerGenerate project scaffolding from paper descriptions (no source code needed)✅ Ready
paper-presenterGenerate Beamer presentations summarizing papers✅ Ready
beamer-skillAcademic Beamer LaTeX presentation lifecycle tool✅ Ready (bundled)
code-reproducerExecute reproduction on remote GPU servers via mcp-ssh✅ Ready
result-analyzerCompare reproduced results against paper figures/metrics. Reports in MD/JSON/Beamer✅ Ready

🚀 Quick Start

git clone https://github.com/orange4664/research-skills.git
cd research-skills

# Use paper-finder
cd paper-finder
pip install -r requirements.txt
python search_paper.py "Attention Is All You Need"

🏗️ Pipeline Architecture

┌──────────────┐    ┌──────────────────┐    ┌──────────────┐
│ paper-finder │───▶│ paper-downloader │───▶│ paper-parser │
│ (search)     │    │ (download)       │    │ (PDF → JSON) │
└──────────────┘    └──────────────────┘    └──────┬───────┘
                                                   │
                    ┌──────────────────┐    ┌───────▼───────┐    ┌──────────────┐
                    │  code-analyzer   │◀───│paper-presenter│───▶│ beamer-skill │
                    │ (AST/ML scoring) │    │ (summarize)   │    │ (LaTeX PPT)  │
                    └────────┬─────────┘    └───────────────┘    └──────────────┘
                             │
          ┌──────────────────┼──────────────────┐
          │                  │                  │
 ┌────────▼─────────┐ ┌──────▼───────┐ ┌────────▼─────────┐
 │  formula2code    │ │ code-writer  │ │ code-reproducer  │
 │ (LaTeX → code)   │ │ (scaffold)   │ │ (train via SSH)  │
 └────────┬─────────┘ └──────┬───────┘ └────────┬─────────┘
          │                  │                  │
          └──────────────────┼──────────────────┘
                    ┌────────▼─────────┐
                    │ result-analyzer  │
                    │ (compare)        │
                    └──────────────────┘

No Source Code? Two new paths:

  • formula2code: Paper equations → PyTorch/NumPy functions
  • code-writer: Paper description → complete project scaffolding (with reference code discovery)

📄 License

MIT License — see LICENSE

🙏 Third-Party

ComponentAuthorLicenseLink
beamer-skillNoi1rMITGitHub
latex2sympy2HuggingFaceMITGitHub
sympytorchpatrick-kidgerApache-2.0GitHub
Reproducibility scoringPapers With CodeML Code Completeness
AST analysis approachPyCG (ICSE'21)Apache-2.0Paper
LaTeX paper writinglatex-paper-skillsMITGitHub

Contributors

orange4664

14 commits

Languages

Python

72.8%

TeX

25.5%

Jinja

1.6%