Cursor 3.8: Improvements to Cursor Automations
CursorCursor released version 3.8 with a significant expansion to its Automations platform, introducing the /automate skill, five new GitHub event triggers, a Slack emoji trigger, and computer use support for cloud automation agents. The /automate skill lets users describe a task in plain language inside a local agent session and Cursor automatically configures the triggers, instructions, and tools. Cloud agents kicked off by automations can now invoke computer use to produce demos or work artifacts as part of their output.
Featured Video
A video we selected to help illustrate this changelog
Sources & Mentions
5 external resources covering this update
The /automate Skill: Natural Language Automation Setup
Cursor 3.8 introduces the /automate skill, a new in-session command that allows developers to create automations without leaving their local agent workflow. Users describe what they want to automate in plain language, and Cursor selects the appropriate triggers, writes the instructions, and configures the necessary tools. Previously, setting up an automation required navigating a separate configuration UI; the /automate skill collapses this into a conversational step.
Slack Emoji Trigger
A new Slack trigger allows teams to activate automations by reacting to any Slack message with a designated emoji. When a matching emoji reaction is detected, Cursor spins up a cloud sandbox and runs the associated automation.
Five New GitHub Triggers
Cursor 3.8 adds five new GitHub event-based triggers: issue comments on non-PR issues; inline comments on PR diffs; PR review submissions; PR review thread status changes (resolved/unresolved); and completed GitHub Actions workflow runs. Two new marketplace templates were added: triaging failed GitHub Actions runs and automatically fixing PR review comments.
Computer Use for Cloud Automation Agents
Cloud agents running inside automations can now invoke the computer use tool to generate demos, screenshots, and other work artifacts. This is enabled by default; users only need to request demo generation in their automation instructions.
Additional Improvements
Automations can now be saved in an incomplete state (preventing data loss during MCP auth setup), open pull requests by default without explicit tool specification, and memory files can be deleted through the UI or via an automation prompt.