Tools for AI-assisted software development.

Supa Magic builds the infrastructure layer that makes AI coding tools actually useful — giving them the right context and giving developers the right environment.

skillbox — A library of individual AI skills.
Architecture guidelines, code patterns, refactoring strategies, testing practices — each extracted from documentation, open-source projects, and real-world conventions. These are the building blocks.

spm — Skill Package Manager.
A CLI that installs skillsets into your project. A skillset is a preconfigured workflow that bundles skills, rules, and integrations for a specific stack. spm install nextjs-fsd and your AI tools know how to work in your project.

runx — One command. Any runtime. Exact version. Zero installs.
A single Rust binary that replaces nvm, pyenv, goenv, and a dozen CI YAML lines. Commit a .runxrc, and everyone on the team — plus CI and Docker — gets the same runtimes.

AI coding assistants are only as good as the context they have. Most projects give them nothing — or a wall of vague instructions.

Supa Magic takes a different approach: skills are small, composable, and versioned. You install exactly what your stack needs. Your AI tools get precise, structured context. Your team gets consistent, reproducible behavior.

Same idea applies to runtimes: a pinned .runxrc means no more "works on my machine" or 12-step onboarding guides.

All projects are open source under the MIT license.