Add llms.txt to repo root for LLM context

Part of #266 — Leonardo AI tooling initiative.

## Summary

Add a `llms.txt` file at the repository root so that LLMs and AI coding assistants can quickly orient themselves to Leonardo's purpose, package structure, and API without reading source files.

## Why

`llms.txt` is a lightweight convention (similar to `robots.txt`) that gives LLMs a structured starting point. Spectrum Design Data uses this same pattern: https://opensource.adobe.com/spectrum-design-data/ai/

## What to include

- Project description and primary use case (contrast-based adaptive color)
- Monorepo structure (`packages/contrast-colors/`, `docs/ui/`)
- Public API summary: `Theme`, `Color`, `BackgroundColor`, `createScale`, `contrast`, `convertColorValue`, `luminance`
- Supported colorspaces for interpolation (LCH, LAB, CAM02, HSL, HSLuv, HSV, RGB, OKLAB, OKLCH)
- Supported output formats (HEX, RGB, HSL, HSV, HSLuv, LAB, LCH, OKLAB, OKLCH, CAM02, CAM02p)
- Contrast methods (`wcag2`, `wcag3`/APCA)
- Links to the npm package and the web app at leonardocolor.io

## Acceptance criteria

- [ ] `llms.txt` exists at repo root
- [ ] Covers all major API exports with brief descriptions
- [ ] Accurate colorspace and output format lists
- [ ] Links to npm, GitHub, and leonardocolor.io

Start at the repository root and review packages/contrast-colors/ and docs/ui/ to confirm Leonardo's purpose, package structure, exports, colorspaces, formats, and contrast methods. Create llms.txt with the requested summaries and links, then verify every acceptance criterion, including the npm, GitHub, and leonardocolor.io links.

Written by the indexing model from the issue text.