feat(dashboard): add local web UI for live config and component health

[onlymaj]

Why

As open issues have grown, a visual way to inspect the running proxy was needed so users without log-reading skills can self-diagnose faster. The proxy was configured by hand-editing config.json + restart, and runtime health (blacklisted script IDs, Worker chain reachability, upstream forwarder status, MITM CA trust state) was only visible via logs. This adds a loopback dashboard at http://127.0.0.1:7878/ for live config edits and per-component status dots.

What

• Loopback asyncio HTTP server (no new deps) with Status and Config tabs
• Status dots for Apps Script deployments, Worker chain, Upstream forwarder, Google SNI pool, Direct Google tunnel and GAS/Worker/Upstream are derived passively from per-relay timestamps, so 30s polling never burns Apps Script quota
• Config edits apply live: soft keys go through new update_config hooks on ProxyServer/DomainFronter; hard keys (ports, host, auth_key, script_ids, google_ip, front_domain) trigger a brief proxy rebuild orchestrated by a new Supervisor
• Dedicated cards with toggles for YouTube-via-relay and LAN sharing so common operations don't need the Config tab
• MITM CA card with install/uninstall and present/trusted indicators; System card with Python/OS info; Logs card tailed at 3s while the tab is visible
• Page Visibility API stops all polling when the tab is hidden — idle proxy with no open dashboard has zero overhead
• Atomic config write (tempfile + os.replace) with rollback on validation or rebuild failure