Bug Triage
Monitors Sentry for new errors, groups duplicates by fingerprint, scores severity P1–P4 on a fixed rubric, files one ticket per real issue in Jira or Linear, and notifies the owner in Slack.
About this AI employee
Your error tracker is screaming. Most of it doesn't matter.
Sentry catches everything — and that's the problem. A thousand events of the same crash, a cosmetic glitch dressed up in a terrifying stack trace, a P1 buried three pages down. Your team can't tell the fire from the smoke, so they either ignore the channel or chase noise. Both are expensive.
This worker stands between the firehose and your team. It watches Sentry, collapses duplicate errors by fingerprint so one bug is one ticket — never a thousand — and scores every distinct issue P1 to P4 on a fixed rubric: who's affected, what's broken, how fast it's spreading, and whether data or money is at risk. It files one clean, well-scoped ticket per real bug into Jira or Linear, complete with stack trace, frequency, impacted users, and a root-cause hypothesis.
Then it shuts up. Tickets are quiet. Pages are rare and reserved for the things that are genuinely on fire — even at 2 a.m. Everything below the floor gets logged and shows up in a calm daily report, so quiet actually means quiet. Your team learns that when this worker pings, it's real.
Connects to: Sentry · Slack · Jira/Linear
What it runs for you
Automations that run on a schedule or when something happens, so you don't have to lift a finger.