lauren

we don't have to live this way. senior engineers have the duty of building guardrails and teaching the next generation of programmers the practices that lead to high quality software. AIs may write most of the code, but you still decide what ships. we should never lose our taste for quality software. i made this analogy a while back: everyone can cook at home but yet restaurants still exist. the floor is rising, but so is the ceiling. the best software will always be made by those who care.

everyone is cooking but who is eating

@brianchew potato with a japanese accent

@poteto yay congrats how do you pronounce it is it po-te-to or pot-eto or po-teto

my twitter name finally matches my github

just shipped web ui bench! measuring taste is hard, so i gave 20 models the same ui component prompts and put every output side by side so you can compare them yourself. let me know which you think is best! built with cursor sdk webuibench.dev

@justinsunyt good news is you can build it your way with the sdk! this was just one of our toy examples

kanban doesn’t make sense for coding agents we tried it 6 months ago. every task just ended up in the “needs review” column

@Tomer_dan @lennysan use our SDK anywhere you wish! cursor.com/docs/sdk/types…

@poteto @lennysan Can we use the Cursor API for hermes or Open Claw?

Narrative violation: Finding it's more fun to work within Cursor than in the native Codex or Claude Code apps. Not a massive difference, but just enough to keep me there. And obviously easier to play with new competing models as they come out. Nice work, Cursor.

@lennysan 1. good feedback, i think we could make that easier 2. it's in Settings > General > Completion Sound! 3. let me see if we can make that clearer. for clarity your Cursor plan is in use here, not your OAI sub

Thanks for asking! 1. Took some real digging to figure out how to give it full permissions, to break out of the sandbox it was in. Much easier to find this in Codex/CC. 2. I bet this is a setting somewhere but I want it ding or notify me somehow when it's ready for my input. 3. I just got this prompt, and I'm not sure if this means I need more credits from my OpenAI account (I have a high plan there), or Cursor covers all of those costs (my assumption). Just hard to know what limit exactly I've reached.

Cursor Security Review is now available for Teams and Enterprise plans. Run two types of always-on agents: 1. Security Reviewer checks every PR for vulnerabilities and leaves comments. 2. Vulnerability Scanner runs scheduled scans of your codebase and posts findings in Slack.

Let's talk about why Cursor's agent harness is so good. There's a misconception that first-party harnesses from the labs will always outperform. For many reasons, that isn't true. There are roughly 6 layers that go into a good agent harness: orchestration, context, routing, transport, state, and execution (tools). Some of those involve careful context engineering, and others look a lot like the traditional craft of building great software. Each of these layers needs to be optimized. And if any one of them is degraded, it can severely impact your experience with the agent. This blog post is mostly focused on context and tools, and there's still so much more to talk about there. We also want to spotlight the other (very crucial!) areas soon. There's a lot more misinformation around the agent harness out there. I'll keep writing about how we build it behind the scenes.

Our agent harness makes models inside Cursor faster, smarter, and more token-efficient. Here's how we test improvements to the harness, monitor and repair degradations, and customize it for different models. cursor.com/blog/continual…

@Joash0x I don’t believe so, but you could make a new slack app and power it using the SDK!

@poteto Is there a way to have the slack cursor app point to a self hosted sdk agent?

set your agents free build on top of the best harness, choose any frontier model, and let agents use a real computer in the cloud. you can just build incredibly ambitious things, and get videos of it all working. will share a cool example i've been cooking up soon

@amix3k cursor /multitask solves this!

Running multiple AI agents in parallel sounds great in theory. In practice, I end up with 10 terminals, 20 PRs, and threads spread across ChatGPT, Claude, and Gemini. Half my time goes into figuring out what’s happening where. It reminds me of adding more monitors. At first, it feels like a productivity boost. Then you realize you’ve mostly created more stress. That’s why I haven’t used multiple monitors in years... Single-tasking for the win. This is mostly just a reminder for myself 😅

cursor sdk launched yesterday! people are already putting cursor agents in places they already work: gmail, chrome, ci, terminal, docs github issues here are 11 projects built in the first day ↓

@BennettBuhner not yet, but it’s in the backlog!

@poteto I skimmed through the docs and didn't see any references to different modes such as ask, plan, and debug. Are those implemented yet?

I've been building with the cursor SDK for a few days now. It's awesome. 🧵 I embedded a cursor agent directly inside Gmail:

@saastrash if you click on About Cursor in the system menu it'll tell you!

@poteto i'm not even sure what version i am using now. how can i tell? lol

in the next version of Cursor 3, you can multitask with a horde of goblins

@juristr let us know what you think!

@poteto I really really need to give Cursor 3 some more love...

We’re introducing the Cursor SDK so you can build agents with the same runtime, harness, and models that power Cursor. Run agents from CI/CD pipelines, create automations for end-to-end workflows, or embed agents directly inside your products.