scitoolkit 0.1.0

A community repository for scientific agentic tools - discover, share, and use AI tools for science

SciToolkit

A Community Repository for Scientific Agentic Tools

SciToolkit is a centralized platform for discovering, sharing, and using scientific tools for AI agents. Think of it as an app store for scientific agentic tools across domains like astrophysics, high-energy physics, quantum computing, neutrino physics, and more.

Vision

Making it as easy as possible for researchers to:

• Create scientific tools for AI agents
• Publish and share tools with the community
• Search and discover tools across scientific domains
• Download and use tools seamlessly
• Contribute to open scientific AI infrastructure

Features (Planned)

• Easy Tool Creation: Build tools using Orchestral AI with universal representation
• MCP Compatible: Full Model Context Protocol support for broad agent compatibility
• Curated Categories: Domain-organized tools (astro, hep, quantum, etc.)
• VSCode Extension: Browse, install, and manage tools directly in your IDE
• Auto-Configuration: One-click setup for Claude Code, Codex, and other agent frameworks
• Web UI: PyPI-like interface for browsing and managing tools
• Version Control: Built-in support for tool versioning and updates

Installation

pip install scitoolkit

Usage (Coming Soon)

For Scripts

# Import tools from specific domains
from scitoolkit.astro import aster
from scitoolkit.hep import heptapod
from scitoolkit.quantum import quantum_toolkit

# Use with your favorite agent framework

For VSCode Extension

1. Install the SciToolkit extension
2. Search for tools by domain or name
3. View metadata, README, and features
4. Install and activate selected tools
5. Auto-configure your agent framework

Included Toolkits

• ASTER (Agentic Science Toolkit for Exoplanet Research) - Tools for exoplanet research
• HEPTAPOD (High Energy Physics Toolkit for Agentic ...) - Tools for high-energy physics
• Quantum Toolkit - Tools for quantum computing research
• Neutrino Toolkit - Tools for neutrino physics research

Architecture

SciToolkit is built on:

• Orchestral AI: Universal tool representation and creation
• MCP (Model Context Protocol): Standard protocol for tool communication
• HTTP/STDIO Servers: Flexible hosting options for tools

Contributing

We welcome contributions! SciToolkit is designed to be community-driven. More details coming soon.

Current Status

🚧 Alpha Stage: This package is in early development. The name is secured on PyPI while we build out the full platform.

License

MIT License

Links

• GitHub: https://github.com/adroman/scitoolkit
• Issues: https://github.com/adroman/scitoolkit/issues