Skip to content

CONTRIBUTING.md

Latest commit

 

History

History
50 lines (35 loc) · 2.31 KB

CONTRIBUTING.md

File metadata and controls

50 lines (35 loc) · 2.31 KB
Raw

Contributing to Quilt

The substrate is live. The paradigm is new. Everything is in motion. That makes this a great time to contribute.

What we need

The biggest leverage points, in rough order:

  1. Adapters. MQTT, Modbus, OPC-UA, GPIO, OpenAI, Anthropic, Ollama. Each one is a 100-300 line module.
  2. Cell type tests. The core has 9 tests. It needs hundreds. Especially for caller-aware caching, cycle detection, async races.
  3. TUI polish. The CLI has a basic text mode. A real tmux-native TUI with cell inspector, dependency overlay, live updates — that's a few days of work.
  4. More examples. Industrial, gaming, agent coordination, scientific computing. Every domain finds a new use.
  5. Web UI. React-based grid with the same affordances as a spreadsheet. The MCP server already does the heavy lifting; the UI is a view.
  6. Documentation. The tutorial is good. The architecture doc is good. We need more recipes, more "how do I" answers, more screenshots.

How to contribute

  1. Open an issue describing what you want to build. We can talk through design before you code.
  2. Fork the repo, push to a branch, open a PR.
  3. PRs need tests. The bar is "the new behavior has a test, and the old tests still pass."
  4. For larger changes, a 1-page design doc helps. We don't want to surprise each other.

Development setup

git clone https://github.com/superinstance/quilt
cd quilt
npm install
npm run test           # all tests
cd packages/core && npm run test   # just core

The runtime is in packages/core. The CLI is in packages/cli. The MCP server is in packages/mcp. Each builds with tsc and tests with vitest.

Coding conventions

  • TypeScript, strict mode, ESM modules
  • Async-first; pure cells are sync, effectful cells are async
  • All cell types implement the same shape: evaluate*(cell, ctx, ...) -> CellValue
  • No any unless you have a really good reason
  • Comments explain why, not what
  • Tests are specs — if a test passes, the behavior is correct

The bigger picture

This project isn't "yet another spreadsheet app." It's a new computational substrate. When you're designing a feature, ask: is this making the substrate more powerful, or is it papering over a missing primitive?

If it's the latter, talk to us. We might be missing the primitive entirely.

Welcome aboard.