Directory
‹ Back to apps

FlareWarden

Check whether any website or service such as Stripe or AWS is down, and monitor your own sites, the services they depend on, cron jobs and status pages with FlareWarden.

What’s inside

1 MCP server · 10 skills
  • MCP serverflarewardenFlareWarden's MCP server: whether a site or service is down (no account needed), and your monitoring, incidents, reports and status pages after OAuth sign-in.
  • Skillchecking-service-statusUse when the user asks whether a website, API or third-party service (Stripe, AWS, GitHub, Slack, Cloudflare, OpenAI, Vercel and others) is down, slow or having an outage, or why something that depends on one is failing ("is X down?", "is it just me?", "checkout is failing, is it Stripe?", "git push hangs for the whole team", "our deploys to us-east-1 time out", "customers in Germany can't load our site"). Reads FlareWarden's public status pages, which put each vendor's own status page beside FlareWarden's checks from several regions, or checks one URL from six regions. No account or sign-in needed. Not for alerts and incidents on the user's own FlareWarden account, which triaging-incidents covers.
  • Skillmanaging-alerts-and-teamUse when the user wants to change who can use their FlareWarden account or where its alerts go, like inviting a teammate at a role, changing a member's role, withdrawing an invitation, seeing who is on the team, getting more or fewer alert emails, pausing them, sending alerts to Slack, Discord, Google Chat, Teams or their own webhook endpoint, testing that alerts arrive, or finding out who changed something in the account ("add Sam to our account as an admin", "make Dana an editor", "stop emailing me about SSL changes", "pause my alerts while I'm on holiday", "send our down alerts to Slack", "add a webhook to our incident endpoint", "did the webhook get the test?", "who paused the API monitor?"). Uses the signed-in FlareWarden MCP tools (list_team, invite_member, change_member_role, update_my_notifications, add_webhook, test_webhook, search_audit_log). For a client who should see only their own sites, or an agency's client sites, use onboarding-client-portfolio. Not for on-call schedules or paging rotations.
  • Skillmonitoring-cron-jobsUse when the user has a scheduled job, cron entry, backup, queue worker, scheduled GitHub Actions workflow, Kubernetes CronJob or systemd timer that could silently stop running, run late, fail or hang, and wants to be told when it does, or asks whether last night's run happened ("our nightly backup stopped for a week and nobody noticed", "alert me if the invoice job doesn't run", "add a heartbeat for this script", "a dead man's switch for the weekly export", "did the backup run last night?"). Plans the schedule and grace period with the signed-in FlareWarden MCP tools (plan_cron_monitor, add_cron_monitor), then adds the ping to the job itself when its code is in reach. Not for writing a cron expression or scheduling a job on its own.
  • Skillmonitoring-third-party-dependenciesUse when the user wants to be alerted when a third-party service that a site they already monitor relies on (a cloud region, CDN, DNS, payment, email, sign-in or search service) has an outage, asks which of its services, components or regions to watch, asks whether a vendor's outage should count as their site being down or only degraded, or asks how a watched dependency has been doing ("alert us when AWS us-east-1 has problems", "our checkout depends on Stripe, watch it too", "which third-party services does acme.com depend on?", "only watch Cloudflare's CDN, not every data centre", "our status page runs on Better Stack, watch that too", "has the Stripe dependency been flapping this week?"). Plans each one with the signed-in FlareWarden MCP tools (plan_dependency, add_dependency, get_monitor_history) before anything is spent. For a site not monitored yet, even when the user names the services it runs on, use setting-up-website-monitoring; for whether a service is down right now use checking-service-status.
  • Skillonboarding-client-portfolioUse when an agency, freelancer or consultant wants to monitor many client websites at once, add a list, spreadsheet or CSV of sites, keep each client's sites in their own group, give a client access to only their own sites, or set up regular reports for each client ("add these 20 client sites", "we only have room for 10 of these 12", "let Bluebird see just their two sites", "we signed three new clients, here are their domains", "how do I organise 50 client sites?"). Uses FlareWarden's bulk plan (one plan for up to 50 URLs, with shared decisions and the total cost against the plan's room), groups and monitor-only invitations through the signed-in FlareWarden MCP tools (plan_site, create_site, invite_member, change_member_access). For one site of the user's own use setting-up-website-monitoring.
  • Skillrunning-status-pagesUse when the user wants a public status page for their customers, or wants to change one they already have, such as its name, address or custom domain (and the DNS record it needs), its colours and logo, what it displays, its search title, or which sites, checks and groups it shows ("give our customers a status page", "put our status page on status.acme.com", "add the API to our status page", "rename the Payments group", "take the old blog off the status page", "why isn't our status page's custom domain working?", "make the status page match our brand colours"). Plans the page from one of the user's sites and publishes it only on the user's yes, then edits it one section or component at a time with the signed-in FlareWarden MCP tools (plan_status_page, create_status_page, update_status_page, add_status_page_component, remove_status_page_component). For posting an incident or maintenance notice on the page use triaging-incidents; for another company's status use checking-service-status.
  • Skillsetting-up-website-monitoringUse when the user wants to start monitoring a website or API, be alerted when it goes down, or set up a new site together with the services it runs on (AWS, Cloudflare, Stripe and others, even when the user lists them) and checks on a page's content ("set up monitoring for acme.com", "tell me if my store goes offline", "we just launched, how will we know if it breaks?", "add our API to FlareWarden", "make sure the Buy button stays on the pricing page", "help me get my first site monitored"). Walks through FlareWarden's plan, decide, create loop with the signed-in FlareWarden MCP tools (plan_site, plan_dependency, create_site), so nothing is created or spent until the user has seen the plan and its cost. For many client sites at once use onboarding-client-portfolio; for cron jobs use monitoring-cron-jobs; for a status page use running-status-pages; for a third-party service on a site already monitored use monitoring-third-party-dependencies; for DNS records and certificates use watching-dns-and-certificates.
  • Skilltriaging-incidentsUse when one of the user's own sites, APIs or cron jobs is down or was down, a FlareWarden alert fired, or the user asks whether an outage was real, what went down last night, what needs attention, why a site keeps alerting, or wants to tell customers on their status page or plan a maintenance window ("we got a down alert at 3am, was it real?", "is anything wrong with our sites?", "FlareWarden says the API is down but it loads for me", "was that outage us or Cloudflare?", "post an update that we're investigating", "schedule maintenance for Saturday night"). Reads the incident's evidence through the signed-in FlareWarden MCP tools (get_account_overview, list_incidents, get_incident, diagnose_site, review_noise) and drafts any status page post or maintenance notice for the user to approve. For a third-party service's status with no account involved use checking-service-status.
  • Skillwatching-dns-and-certificatesUse when the user wants to know when DNS records or TLS certificates for their own domain change, stop resolving or are about to expire, such as MX, SPF or DMARC records, a CNAME like www pointing at Vercel, an SSL certificate nearing expiry, a certificate issued for their domain by an authority they don't use, or a subdomain's certificate, or asks what a DNS or certificate alert from FlareWarden means ("tell me if our MX records change", "alert me if www stops pointing at Vercel", "watch _dmarc for changes", "warn me well before our certificate runs out", "was a certificate issued for our domain that we didn't ask for?", "watch the certificate on api.acme.com too", "what DNS records does acme.com have?"). Uses the signed-in FlareWarden MCP tools (plan_dns_watch, add_dns_watch, get_site, track_subdomain_certificate, update_my_notifications). For whether a site is down right now use triaging-incidents; for many client sites at once use onboarding-client-portfolio.
  • Skillwriting-uptime-reportsUse when the user wants an uptime, availability, SLA or incident report for a website, an agency's client or the whole account over a day, week, month or quarter, or asks what uptime or SLA their own sites actually achieved, how much downtime there was, whether an uptime target was met, how fast sites recovered, or how response times compare ("write October's report for Bluebird", "how did we do against our SLA last quarter?", "did we hit 99.9% this quarter?", "uptime numbers for the board deck", "send each client a monthly report", "what was our mean time to recover?"). Builds the report from FlareWarden's measured figures through the signed-in FlareWarden MCP tools (get_report, list_incidents, get_incident) and writes the words around them without changing a number. Not for whether something is down right now.