Cursor Automations now support event-based triggers for Sentry. Set up agents that automatically respond to new issues, investigate root causes, open PRs with fixes, and post summaries to Slack.

Start building a Sentry automation with a template from our marketplace:

Is it made with @screenstudio ?

For those who like to cook, here's a new recipe 😏

@JonasBadalic Sentry 🤝 Cursor

the failure mode i hit running this manually: 5 errors fire at once, you get 5 PRs each fixing a symptom, all the same root cause. whatever dedup step lives before the agent trigger is doing more work than the agent itself

The root cause investigation piece is huge. Most debugging still happens in reactive panic mode when production breaks. Having an agent that can trace through logs, identify patterns, and propose actual fixes while you sleep changes everything about incident response.

Been building agents for debugging workflows - the hardest part is getting them to stop at "investigate" instead of immediately trying to fix everything.

that is nice feature responding to incidents like these is critical

Will this cost extra money if the agent runs a lot on my issues?

Cursor Automations with Sentry event triggers is exactly the kind of autonomous loop developers have been waiting for. New issue → root cause analysis → PR opened → Slack summary. Incident response just went from reactive panic to silent, reliable fixes.

Sentry hook → auto PR in minutes? my on-call rotation just got deleted.

@_smontlouis 👀

claude code ships routines now, cursor ships sentry automations the same time. the speed of this feature war is getting ridiculous

Now Sentry errors get automatically investigated, fixed, and posted to Slack

Having your own software it’s quite noticeable in shipping speed. Glad you guys got rid of the fork finally

if claude isn’t fixed in the next month im going back to cursor i said it

The Sentry to Slack loop is the holy grail for incident triage. Having an agent investigate root cause before anyone gets paged is where this gets interesting. Especially for noisy, low severity errors that nobody reviews until there are 200 of them.

Nice step. The interesting part now is not just firing the loop from Sentry. It is making each run legible enough that a team can trust it in production: what triggered it, what context it used, what changed, and how to stop or roll back a bad fix. Fast automations are easy to demo. Inspectable ones are easier to keep.

can’t wait for the agent to accidentally delete the repo at 3am.

Agents that open PRs from Sentry issues automatically?? This is genuinely game changing!

u guys in the same room?

set once and it just handles on-call chaos while I sleep? this is the ‘agents actually ship’ moment I’ve been waiting for

An AI agent that automatically catches a Sentry error, figures out why it happened, and opens a PR to fix it is a genuinely big workflow shift. Less time fire fighting, more time building.

Event-driven agents responding to real production signals is where agent reliability actually matters. The agents that win aren't the smartest — they're the ones with constrained scope, clear triggers, and verified outputs. This is the pattern we keep seeing in our benchmarks.

HUGE

可以关闭自动创建 subagent 功能吗

triggers still feel too limiting

Event-based Sentry triggers for automated debugging agents is exactly the kind of integration devs need. Auto-investigating root causes and opening PRs with fixes — that's moving from 'AI assistant' to 'AI teammate' territory.

More autonomous way of handling errors

less debugging pain now

Nice! Now we just need GitHub Issue triggers

Atlassian wen?

can we post to discord too?

Just waiting for macOS images in cloud agents instead of linux images.

When datadog?