just wanna say… not sure what’s happening this update, but the composer option next to where the chat window is feels like an absolute beast with agent mode enabled…
Im working on an app with a go backend repo and a react-router 7 repo with simple react bootstrap (which im using the full framework so essentially remix 3). Even though LLMs would’t have much training on this as new, it just knows what to do based on the other patterns through the codebase and files…
The way i work is to have a separate repo for the api and then the front end and have 2 cursor windows open… if i need to tell the frontend about API calls from the backend, I just paste in the part from my postman collection.
and if it can’t fix on a lint, I just provide a page of documentation and it gets it right.
not sure what you’ve done this update and I never normally post on these forums, but bravo guys, this certainly feels like a massive leap forward…
Another thing I like is when you right click and error and select fix in composer it adds it to the current composer context.
So now, i no longer user chat at all, and stick with composer, but the one next to the chat.
what sometimes I need to do is the recalculating of index doesn’t catchup quite quick enough before I send my next message, so i go in and manually do it as it usually just takes a few seconds, and then i’ll resend…
Edit:
just to note I have no custom rules either… i just tell it to follow existing patterns in other files. ymmv
