feat(bot): add a /status command that names the failing layer

[matthew-demidoff]

Summary

• /status in the alert chat (or from the admin in a private chat) replies with a layer-by-layer breakdown and a verdict: UP, DEGRADED or DOWN, plus the layer to blame (host app, worker, tunnel, Cloudflare edge) or a note that a remaining failure is on the caller's side
• the bot probes the app's readiness from inside, cloudflared's /ready, the external monitor's recorded state, and Cloudflare's public status page, in parallel with bounded timeouts; the request path is walked from the origin outwards and the first broken layer wins
• the monitor Worker gains a token-gated GET /status returning its checks and heartbeats with its own clock, so heartbeat ages do not depend on the host clock
• node index.js --status prints the same report from a shell; the command is ignored outside the gated chats
• compose passes ALERT_TELEGRAM_CHAT_ID, MONITOR_STATUS_URL and CLOUDFLARED_READY_URL to the bot; CI builds the bot image and imports the new module; docs and runbook updated, including a --no-deps note for rebuilding the worker or bot without recreating the app

Verification

• unit tests cover the command gate and every verdict branch (host down, unreachable app, tunnel without connections, edge failure with a Cloudflare incident, stale worker, unreachable or unconfigured monitor, shell run)
• deployed: the Worker route answers 404 without, with a wrong token, or on POST, and 200 JSON with the right token; docker compose exec bot node index.js --status in production printed a complete report with all layers ok and Cloudflare's current minor incident shown