Incident Watch
Watches Sentry and Vercel for error spikes and bad deploys, opens incidents, summarizes blast radius, alerts and pages your Slack channel, keeps a live timeline, and drafts a blameless postmortem when it's over.
About this agent
Cut the time between "something's broken" and "the right person is on it" to seconds.
Small teams can't staff a 24/7 incident commander. Incident Watch is the next best thing: it never sleeps, it watches Sentry and Vercel every five minutes, and the instant errors spike or a deploy goes bad, it opens an incident, figures out the blast radius, and pages your channel — with the suspected commit and author already pinned.
While the incident runs, it keeps a clean, timestamped timeline so nobody has to reconstruct "what happened when" from memory. When it's over, it drafts a blameless postmortem with real impact numbers and concrete action items, ready for a human to review and sign off.
It is fast, factual, and disciplined: it never touches production, never cries wolf over a single error, and never points a finger at a person.
Connects to: Sentry · Vercel · Slack
What it runs for you
Automations that run on a schedule or when something happens, so you don't have to lift a finger.