Cursor automates Sentry workflows

Cursor is pushing software alerts closer to auto-repair: its new automations can now start from a Sentry issue and hand work to a cloud coding agent. (cursor.com) (docs.sentry.io) Cursor introduced Automations on March 5, 2026 as always-on agents that run on schedules or event triggers from tools including Slack, GitHub, Linear, PagerDuty, and custom webhooks. When triggered, the agent starts a cloud sandbox, uses configured models and Model Context Protocol tools, and can verify its own output. (cursor.com 1) (cursor.com 2) Sentry already had a handoff into Cursor before this latest demo. Sentry said on November 20, 2025 that Seer, its artificial intelligence debugging agent, could send issue context and root-cause analysis to a Cursor Cloud Agent to write a fix and open a pull request. (blog.sentry.io) (docs.sentry.io) For developers who do not live in these tools, Sentry is the system that collects crashes, traces, logs, and other production signals after software breaks. Seer is Sentry’s built-in debugging agent, which can analyze incoming issues, generate code fixes, and delegate work to outside coding agents such as Cursor. (docs.sentry.io 1) (docs.sentry.io 2) The new piece in Cursor’s video is the trigger point. Instead of waiting for an engineer to copy an error into an editor, a monitoring event can kick off an investigation, draft a pull request, and post a Slack summary as part of the same workflow. (cursor.com) (docs.sentry.io) That fits the way both companies have been expanding around Model Context Protocol, the standard that lets an artificial intelligence tool call outside systems for context and actions. Sentry’s MCP server gives coding tools direct access to issues, projects, and Seer analysis, while Cursor says its automations use configured MCPs inside the cloud agent run. (docs.sentry.io) (cursor.com) Sentry’s documentation says teams can also trigger Cursor Cloud Agents automatically through Seer Automation by selecting Cursor as the coding agent in project settings. Cursor’s own examples for Automations already include background review, security checks, and incident-response workflows that post findings into Slack. (docs.sentry.io) (cursor.com) The pitch from both sides is less about chat inside an editor and more about wiring software operations into an automatic loop. In that loop, the alert is no longer just a notification for a human; it is the start of a job that can investigate, code, and report back. (sentry.io) (cursor.com)