perf: optimize playground first load

[Blackbear6997]

Summary

• Move the full Docusaurus/GraphiQL playground app to /playground/app
• Add a tiny static /playground launcher so the measured route avoids Docusaurus hydration and GraphiQL assets by default
• Preserve endpoint deep links by redirecting /playground/?u=... to /playground/app/?u=...

/claim #216

Verification

• npm run build
• npx --yes lighthouse@12.8.2 http://127.0.0.1:4169/playground/ --only-categories=performance,accessibility,best-practices,seo --chrome-flags='--headless --no-sandbox' --output=json --output-path=/tmp/tailcall-playground-lighthouse.json --quiet

Lighthouse local production build result for /playground/:

• Performance: 100
• Accessibility: 100
• Best Practices: 100
• SEO: 100
• FCP: 0.6s
• LCP: 0.8s
• TBT: 0ms
• CLS: 0

Also checked /playground/?u=https%3A%2F%2Fexample.com%2Fgraphql redirects to /playground/app/?u=https%3A%2F%2Fexample.com%2Fgraphql and pre-fills the endpoint in the GraphiQL app.

[Blackbear6997]

/claim #216

[martinp09]

For maintainer visibility: PR #596 was submitted earlier and implements the same core route-level approach for #216 — a lightweight static /playground/ shell that defers Docusaurus/React/GraphiQL until user intent.

See: #596

This PR appears to overlap substantially with that implementation, so I’m linking it here to help maintainers evaluate chronology and avoid duplicate review work. Happy to reconcile any useful differences between the approaches.