DEEP_DIVE_STORY.LNK

Cursor launched Automations, a policy-driven system that triggers coding agents on commits, Slack messages, or schedules and loops humans in only when needed.
Automations shift work from prompt-and-monitor to always-on workflows that watch your repository, kick off scans, and route results to humans via PRs or Slack when judgment is required (TechCrunch, FindArticles). Teams can define triggers for pre-merge static checks, post-merge tests, refactors, and dependency safety, then dispatch specialized agents that report back with actionable diffs and comments.
The feature builds on Cursor’s long-running Bugbot, expanding it into layered reviews and security sweeps with higher token budgets to catch subtler defects—an approach leaders say has surfaced issues linters and CI often miss (TechCrunch, FindArticles). Expect multi-stage passes that escalate from quick scans to deeper traces and targeted test generation.
Early community reports flag some rough edges unrelated to Automations—like missing settings after a recent update and overly noisy lint flags—so plan for careful policy tuning and review routing as you pilot (forum thread 1, forum thread 2).