Triage New Sentry Issues

The Idea

A Sentry Issue created trigger starts a run for each new issue. The agent gets the issue's title, link, and event payload in its prompt, finds the code in the stack trace, reproduces the bug in a test, and opens a draft pull request.

Steps

1. Connect Sentry under Integrations.
2. Go to Automations → New automation. Pick the project for the repository the Sentry project reports from, a device that stays online, and an agent.
3. Under Triggers, click Add Trigger → Sentry → Issue created. Pick the Sentry project and set the level to Error and Fatal, so warnings don't start runs. Press Save.
4. Write a prompt that proves the bug before it fixes it:

A new Sentry issue is in the event above. Find the code in its stack trace.
Write a failing test that reproduces the error. If you can't reproduce it,
stop and write what you found to TRIAGE.md.
If you can, fix it, run the test suite, and open a draft pull request with
`gh pr create --draft`. Link the Sentry issue in the description, and say
what caused the error and how the test proves the fix.
If a pull request for this Sentry issue already exists, stop.

5. Open the run from Run History, read TRIAGE.md or the draft pull request, and mark it ready for review when you agree.

Tips

• Reproduce first: a failing test shows that the fix works. Without one, you cannot tell.
• Draft only: --draft keeps CI and reviewers out until you have looked at it.
• Duplicates: the same event can be delivered twice. The "already exists" check stops the second run.

Variations

• Regressions: use Issue unresolved to catch a bug that came back after it was marked resolved.
• Assign to the agent: use Issue assigned so only the issues someone hands over start a run.