Introducing the Agent SDK

Cursor has released the generally available Agent SDK, a framework for building, testing, and deploying code-defined software agents that automate recurring development tasks such as pull-request reviews, compliance workflows, and incident investigations. The SDK enables teams to tailor agents to their own processes, connect them to events from platforms including GitHub, GitLab, and Bitbucket, and encode policies that can be version-controlled and reviewed alongside application code. Agents can use specialized subagents, parallelize work, and rely on built-in evaluations and event replay to validate critical workflows, with examples including Cursor’s knowledge-base pull-request reviewer and the PGA Tour’s software development factory. Agent projects organize prompts in Markdown, tools as TypeScript functions, external integrations, and tests in a small software-project structure, and can be deployed to Cursor’s infrastructure through a single command or run independently as containerized Node processes.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.