Cursor Automations Go Live

Cursor is launching Cursor Automations, a new feature designed to build 'always-on' agents. These agents can be scheduled to run or triggered by events like Slack messages, new Linear issues, merged GitHub pull requests, or PagerDuty alerts. Custom events are also supported via webhooks.

This move aims to address the widening gap between the rapid pace of code generation enabled by AI agents and the slower progress in code review, monitoring, and maintenance. As detailed in AI Agents Usher in Software Development's Third Era, scaling these essential but less glamorous parts of the lifecycle is critical.

Scaling the Software Engineering Pipeline

Cursor Automations works by spinning up cloud-based sandbox environments for agents. These agents execute instructions using configured models and tools, verifying their own output. A memory tool allows them to learn from past runs, improving over time.

The company has identified two primary categories for these automations: review and monitoring, and chores.

Review and Monitoring

Automations are positioned as a powerful tool for code review, capable of catching everything from minor style issues to significant security vulnerabilities and performance regressions. Cursor’s own Bugbot, a long-standing automation, exemplifies this, having identified millions of bugs since its inception.

Cursor highlights three specific use cases:

• Security Review: Triggered on every push to the main branch, this automation audits code for vulnerabilities without blocking PRs, escalating high-risk findings to Slack.
• Agentic Codeowners: On PR open or push, this automation assesses risk and auto-approves low-risk changes or assigns reviewers for higher-risk ones, logging decisions to Notion.
• Incident Response: When a PagerDuty incident occurs, an agent investigates logs and recent code changes, then proposes a fix via a PR in Slack.

Chores and Routine Tasks

Beyond critical reviews, automations can handle routine tasks and knowledge work that span multiple tools. Examples include generating weekly summaries of repository changes, identifying and adding test coverage to recently merged code, and triaging bug reports by checking for duplicates, creating issues, and attempting fixes.

Real-World Adoption

Early adopters are already integrating automations into their workflows. At Rippling, an employee uses a personal assistant automation that consolidates meeting notes, action items, and other information into a dashboard by cross-referencing Slack, GitHub, and Jira. Rippling also leverages automations for incident triage and status reporting.

Tal Peretz, Co-founder of Runlayer, noted the impact on their development process: "We built our software factory using Cursor Automations with Runlayer MCP and plugins. We move faster than teams five times our size because our agents have the right tools, the right context, and the right guardrails."

Cursor Automations aims to transform the development process into a 'software factory' where agents continuously monitor and improve codebases.