Software teams increasingly rely on agents defined as code to handle recurring work across the development lifecycle. These agents connect to the tools a team already uses and act when relevant events occur, such as reviewing a pull request, executing a compliance workflow, or investigating an incident. The most useful agents are tailored to how a particular team operates. Until now, building one generally meant adopting a product designed around someone else's workflow or starting from scratch.
Today, the company is releasing the Agent SDK, the framework it uses to build its own managed agents, including Bugbot, Security Reviewer, and Rollouts. The SDK lets developers define agents in code and manage them like any other software project. Teams can start from a template or point a coding agent at the documentation. When an agent is ready, it can be deployed to Cursor infrastructure with a single command.
The SDK is designed to match each team's workflow. Agents can respond to pull request events from GitHub, GitLab, or Bitbucket, assess the risk of a change, route it to the right team, and approve changes that meet policy. A Cursor team built kb-reviewer in about a day to check every pull request against its knowledge base of lessons learned, flagging contradictions and citing the relevant note. For complex jobs, agents can hand tasks to specialized subagents or divide work among them in parallel. Evals and event replay are supported out of the box.
The PGA Tour uses the Agent SDK to power its software development factory. Michael Gyarmathy, Principal Engineer at PGA America, said the SDK provides the flexibility needed to build agents that drive the development process, while built-in evals allow the team to iterate quickly and optimize capabilities reliably. Agents are organized like small software projects, with Markdown prompts, TypeScript tools, and evals, so changes move through the same version control and review process as the rest of the codebase.
Source: Hacker News · Summarized by HeadlinesBriefing