Cursor launches an Agent SDK for teams building custom agents

Cursor released its Agent SDK in general availability on October 20, letting software teams define agents in code, test them with evals and event replay, and deploy them to Cursor's infrastructure with one command or run them on their own. The SDK is the same framework Cursor uses for its managed agents, including Bugbot, Security Reviewer and Rollouts, and lets teams trigger agents from workplace events. Cursor said teams remain responsible for the policies and evaluations that make agents dependable.

Cursor launches an Agent SDK for teams building custom agents Cursor's October 20th release lets software teams define agents in code, test them with evals and event replay, and deploy them on Cursor's infrastructure or their own. By Ryan Merket https://runtimewire.com/author/ryan-merket · Published Primary source: Hacker News https://cursor.com/blog/agent-sdk Why it matters Cursor is extending its coding-agent platform into the workflows around software delivery. The SDK gives teams a code-based way to customize those workflows, while leaving them responsible for the policies and evaluations that make agents dependable. Cursor's October 20th product post https://cursor.com/blog/agent-sdk?ref=runtimewire says the Agent SDK was generally available that day; the date also appears in Cursor's changelog https://cursor.com/blog?ref=runtimewire . The framework Cursor uses for managed agents such as Bugbot, Security Reviewer and Rollouts lets software teams build agents around their own processes, trigger them from workplace events and deploy them to Cursor infrastructure with one command.… When you purchase through links in our articles, we may earn a small commission https://runtimewire.com/affiliate-monetization-standards . This doesn’t affect our editorial independence.