scientific-agent-skills

Scientific Agent Skills

A comprehensive collection of 72 Agent Skills for scientific computing, research workflows, and data analysis. These skills automatically enhance AI coding assistants (like Cursor, Claude Code, and others) with deep domain knowledge across the entire scientific Python ecosystem.

What Are Agent Skills?

Agent Skills are contextual knowledge modules that automatically load when an AI assistant detects relevant topics in your conversation. Instead of generic coding help, you get expert-level guidance tailored to specific scientific domains.

How It Works

1. You start coding with scientific libraries (NumPy, PyTorch, scikit-learn, etc.)
2. Skills auto-load based on semantic matching of your code and questions
3. AI gets expert knowledge from detailed skill documentation
4. You receive better assistance with best practices, common patterns, and domain-specific solutions

Example

When you describe your problem:

"I need to process a large array of scientific data efficiently. How do I optimize NumPy operations for performance?"

Or when you ask:

"What's the best way to handle missing values in pandas before training a machine learning model?"

The relevant skills (numpy or pandas-performance) automatically load, providing the AI with:

• 1,361 lines of expert documentation (for NumPy)
• Best practices for array operations
• Common pitfalls and solutions
• Performance optimization techniques
• Real-world patterns and examples

The AI then gives you expert-level guidance tailored to your specific problem, not generic coding help.

What Problems Can This Solve?

🔬 Scientific Computing & Research

• Numerical Computing: NumPy, SciPy, SymPy, JAX for mathematical computations
• Physics & Chemistry: Quantum computing (Qiskit, PennyLane), molecular dynamics (MDAnalysis), quantum chemistry (PySCF)
• Astronomy & Astrophysics: Astropy, SunPy, Photutils for astronomical data analysis
• Bioinformatics: Biopython, scikit-bio, Scanpy for biological data processing

🤖 Machine Learning & AI

• Deep Learning: PyTorch, TensorFlow for neural networks
• Classical ML: scikit-learn, XGBoost, LightGBM for traditional algorithms
• NLP: Transformers, spaCy, NLTK for natural language processing
• Graph Neural Networks: PyTorch Geometric for graph-structured data

📊 Data Analysis & Visualization

• Data Processing: Pandas, Polars, XArray for tabular and multidimensional data
• Visualization: Matplotlib, Seaborn, Plotly for creating publication-quality plots
• Statistical Analysis: Statsmodels, PyMC, NumPyro for Bayesian inference

🗺️ Geospatial & Earth Science

• Geospatial Analysis: GeoPandas, Shapely, Rasterio for geographic data
• Remote Sensing: Rasterio for satellite imagery processing
• Coordinate Systems: PyProj for map projections

⚙️ Optimization & Simulation

• Mathematical Optimization: OR-Tools, Pyomo for linear/nonlinear programming
• Bayesian Optimization: Ax Platform for experiment design
• Simulation: SimPy for discrete event simulation
• Parallel Computing: Dask for distributed computing

🔬 Specialized Domains

• Signal Processing: MNE for neurophysiological data (EEG/MEG)
• Medical Imaging: PyDICOM for DICOM file processing
• Time Series: sktime, tsfresh for temporal data analysis
• Causal Inference: DoWhy for causal analysis beyond correlation

What Systems Can You Build?

With these skills, you can build:

Research & Analysis Systems

• Scientific Data Pipelines: Automated processing of experimental data
• Statistical Analysis Tools: Bayesian inference, hypothesis testing, causal analysis
• Visualization Dashboards: Interactive plots for research publications
• Simulation Frameworks: Physics simulations, molecular dynamics, quantum systems

Machine Learning Systems

• Research Prototypes: Rapid prototyping of new ML architectures
• Production ML Pipelines: End-to-end training and deployment workflows
• Explainable AI Tools: Model interpretation and feature importance analysis
• AutoML Systems: Automated hyperparameter optimization with Ax

Scientific Computing Platforms

• Computational Chemistry Tools: Quantum chemistry calculations, molecular modeling
• Astronomical Data Processors: Image analysis, catalog matching, light curve analysis
• Bioinformatics Workflows: Sequence analysis, protein structure prediction
• Geospatial Analytics: GIS analysis, remote sensing applications

Domain-Specific Applications

• Medical Imaging Analysis: DICOM processing, image segmentation
• Neuroscience Tools: EEG/MEG signal processing and analysis
• Materials Science: Crystal structure analysis, property prediction
• Climate Science: Time series analysis, spatial data processing

Installation

For Cursor IDE