Recommended starting points
- Sentry error triage
- GitHub issue response
- Linear ticket first pass
What success looks like
- Faster time‑to‑fix for recurring issues
- Cleaner handoffs between support and engineering
- Clear visibility into what the agent changed and why