cursor-sdk

Integrate the Cursor TypeScript SDK for building and running agents in code, CI, and automation.

2.6k|204|Updated Jan 23, 2026
One-click install
npx skills add https://github.com/cursor/plugins --skill cursor-sdk
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cursor-sdk
Source: https://github.com/cursor/plugins/tree/main/cursor-sdk/skills/cursor-sdk
Command: npx skills add https://github.com/cursor/plugins --skill cursor-sdk

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

The Cursor SDK skill helps developers bootstrap and reason about integrating Cursor's TypeScript SDK into code, CI pipelines, and automation tasks, making it easier to design durable agent integrations and avoid common runtime pitfalls.

Core Features & Use Cases

  • Guidance on choosing between local and cloud runtimes and configuring MCP servers.
  • Patterns for streaming vs waiting, error handling, resume workflows, and artifact management.
  • Practical integration examples like one-shot prompts, durable agents, and parallel fan-out across repos.

Quick Start

Install the Cursor SDK, create an Agent with an API key, choose a runtime (local or cloud), and issue a prompt.

Frequently Asked Questions about cursor-sdk

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I integrate Cursor SDK into TypeScript automation workflows?

To integrate the Cursor SDK into TypeScript automation, install the SDK, create an agent with an explicit API key, and choose a local or cloud runtime. You can then issue one-shot prompts or build durable agents for CI pipelines and automation tasks.

What is the difference between local and cloud runtimes for Cursor agents?

Local and cloud runtimes determine where your Cursor agents execute. The SDK provides guidance on choosing between them based on your execution environment, configuring MCP servers accordingly, and managing agent lifecycles across different runtime contexts.

How do I handle streaming responses and durable agent lifecycles in Cursor SDK?

The Cursor SDK supports both streaming and wait patterns for responses. For durable agent lifecycles, it provides resume workflows and artifact management capabilities, allowing you to handle long-running processes and recover state across execution cycles.

Can I run parallel agent fan-out across multiple repositories with Cursor SDK?

Yes, the Cursor SDK supports practical integration examples including parallel fan-out across repositories. You can configure multiple agents to operate concurrently across different codebases using explicit API keys and robust error handling patterns.

What are the limitations when configuring MCP servers with Cursor SDK?

When configuring MCP servers with the Cursor SDK, you need to consider runtime compatibility between local and cloud environments. The SDK provides configuration guidance but requires explicit API key usage and proper error handling to avoid common runtime pitfalls.