Rewrite root README with a complete, accurate example catalog

[NikhilShahi]

The root README was a generic placeholder: wrong clone URL (your-org/...), Node 16+, claims of tests and a LICENSE file that did not exist, and no index of the 46 apps.

This replaces it with a README that matches what is actually in the repo.

What changed

• Correct clone URL: github.com/hyperbrowserai/hyperbrowser-app-examples
• Table of contents, prerequisites, and a standard Next.js quick start
• All 46 standalone apps, grouped by use case, with a one-line description and typical API keys
• “How to pick an example” table (Live View, Computer Use, sandboxes, swarms, DESIGN.md, etc.)
• Run notes for the outliers: ChurnHunter CLI, Hyperplex (Docker + worker), Agent Web Index harness, UI-entered keys, o5-automation preflight
• Honest env-file conventions (templates are inconsistent; many READMEs reference a missing .env.example)
• Hyperbrowser capability map (scrape, Live View, Computer Use, sandboxes, volumes)
• Deployment caveats (serverless timeouts, cost, local public/runs/ disk)
• Real support links only (docs, dashboard, Node SDK, issues, @hyperbrowser)
• Root MIT LICENSE so the license section and badge are valid (three apps already shipped MIT)

What I did not invent

• No Discord/community link (none exists in-repo)
• No claim that every app is production-ready or has tests
• Node guidance is 20.9+ recommended (Next.js 16); 18.18+ for older Next 14/15 apps

Each app still has its own README for setup details. This file is the index and the shared getting-started guide.