instant-mobile-notifications

Install

$ curl -fsSL -L https://tetsukod.ai/api/skills/instant-mobile-notifications/download | tar -xz -C ~/.claude/skills/

Drops the `instant-mobile-notifications/` skill into `~/.agent/skills/`. Once it lands, `/instant-mobile-notifications` is available in any Claude Code session run from a project with Grafana reachable.

What it does

Wire your Grafana alerts to your phone via ntfy push notifications — one interactive run that provisions a webhook contact point, a default notification policy, and a starter LogQL alert rule, all as Grafana provisioning YAML the project owns. The skill smoke-tests ntfy reachability with a direct test POST to your phone before touching Grafana, because wiring a broken push channel is the worst failure mode this kind of setup has.

What you get:

• A Grafana webhook contact point using ntfy's built-in ?template=grafana parser — Grafana's native JSON parses on the ntfy side, no custom message-templating wrangling in Grafana
• A default notification policy that groups firing alerts by alertname and service so a burst arrives as one push, not thirty
• Starter LogQL rules (application-errors, service-silence) you can extend by writing more YAML — Grafana picks them up on container restart
• A pre-Grafana smoke test that proves the push channel works before any alerting wires touch Grafana
• Per-rule overrides via ntfy_* labels (title, priority, tags) so you can route critical alerts differently from informational pings

What's inside

The package ships every YAML and reference doc the skill installs into your project — a Grafana webhook contact point template, a default notification policy, a starter rule template with two candidates pre-shaped (application-errors, service-silence), and an ntfy-templating reference explaining the wire and how to customize per-rule pushes. The orchestrator in SKILL.md runs the interview, the phone-side smoke test, the YAML provisioning, and the Grafana reload.

When to reach for it

Reach for this skill when you have Grafana running and want firing alerts to ping your phone instead of sitting in a dashboard waiting for someone to notice. It is the right fit for solo founders and small teams who already use ntfy for agent pings — the wire is the same channel, so the phone subscription you already have just picks up another voice. If you ran /instant-observability first, this skill reads its .env.local and inherits the Grafana credentials; if Grafana is somewhere else, you paste the URL and key at the interview step.

Skip it for email or Slack as the alert routing target — those are different contact points with their own quirks. Skip it for projects without Grafana yet — run /instant-observability first. And skip it for production-grade incident management with on-call rotations, paging escalations, and acknowledge-or-page-the-next-person flows — this is the phone-buzzes-when-something-fires shape, not a PagerDuty replacement.