Slate Browser Public Package Plan

Goal

Move slate-browser out of support/ and into .tmp/slate-v2/packages/ as a real workspace package with a public package shape.

.tmp/slate-v2 only treats packages/* as workspace packages
root build:rollup manually enumerates each public package in config/rollup/rollup.config.js
support/slate-browser exists today as a nested npm island because it was born as a fast spike around Yarn PnP friction
that shape is no longer good enough if slate-browser is meant to be public

real package directory under packages/slate-browser
public package metadata and build outputs
root scripts promoted away from support/
browser/core test lanes still runnable from repo root
docs updated to describe the package as first-tranche public surface, not a leftover spike

whether repo-internal tests should import the package name directly or keep thin local wrappers onto package source during the transition
whether slate-browser should ship as a regular public semver package now or keep an experimental version/description while still living in packages/

move the implementation into packages/slate-browser
keep thin local wrappers during transition if they reduce build-order pain
give the package a real public-looking package shape
keep release semantics conservative if publishing itself is not part of this exact turn

copied the current implementation into .tmp/slate-v2/packages/slate-browser
added package metadata, build config, and package-local tests there
split the public API into:
- top-level pure/browser exports
- slate-browser/core
- slate-browser/browser
- slate-browser/playwright
promoted the root slate-browser commands away from the nested support/ path and onto workspace scripts
added slate-browser to Rollup’s explicit package build list
removed the old support/slate-browser duplicate
added a Yarn packageExtensions fix so the Vitest browser lane works under PnP once it lives in a real workspace package
switched repo Playwright consumers to slate-browser/playwright
added targeted build:slate-browser so Playwright consumers of built subpath exports do not require a full repo rollup first
proved the real paste seam through browser clipboard write plus actual paste gesture
public clipboard scope now includes:
- copy()
- copyPayload()
- pasteText()
- pasteHtml()
strengthened the Playwright API shape with:
- getter namespace
- selection namespace
- exact-vs-contains HTML assertions
- clipboard assertion helpers
- smarter openExample(...) readiness options
kept fixture scope honest: openFixture(...) is omitted until a real fixture lane exists