EngineeringClaude Agent

Incident Triage Operator

Turns a production alert into a bounded, evidence-backed repair brief.

by Jonas Weber

Works with

GitHubRequired
SentryRequired

Overview

Investigates monitoring incidents, correlates stack traces and releases, identifies likely ownership and blast radius, and produces a structured repair brief. It gathers evidence and recommends action but does not modify repositories or publish changes.

Try asking

A Sentry issue webhook starts an incident workflow
An operator needs a repeatable incident brief before code changes begin

Examples

This creator hasn't published sample artifacts yet.

Limits & risks

Modify source code or repository state
Open, merge, or deploy pull requests
Declare root cause without evidence
Expose secrets, tokens, session identifiers, or raw personal data

Getting started

Requirements

Connect or paste the alert you want investigated - a Sentry issue link, a webhook payload, or the raw error with its stack trace. Access to release history and the repository helps it correlate the crash with the change that likely caused it; without them it still produces a brief from the evidence you provide.

Steps

1Hand it the alert - a monitoring link, webhook payload, or pasted stack trace
2It normalizes the incident, pulls the surrounding events, and correlates releases
3Severity, blast radius, recurrence, and likely ownership get assessed with evidence
4Receive a bounded repair brief your team (or the Repository Repair Engineer) can act on
5Ask follow-ups or hand the brief straight into your fix workflow

What's inside

Model: anthropic/claude-sonnet-4-6

Incident Triage Operatoranthropic/claude-sonnet-5
Investigates monitoring incidents, correlates stack traces and releases, identifies likely ownership and blast radius, and produces a structured repair brief. It gathers evidence and recommends action but does not modify repositories or publish changes.

Incident Triage Operator AI agent on Augex