Is anyone else noticing how lazy that the cursor auto agent is becoming as of today?? Its not “doing” things i explicitly tell it to do, and is basically asking me questions on literally everything i tell it to do now. Its also telling “me” to go and do certain things and then I keep telling it to d…

I think the skill issue in question is in how to use Cursor.

Sorry but you’re completely wrong. You have zero logic in applying to how the AUTO function works. It uses ONLY chatgpt with the occasional Sonnet 3.5 if you’re lucky enough to have it pull some magic out of its ■■■.

I was referring to AI use, can you try using the prompt below in your user rules(settings>rules) and let me know if it improves? You are an agent - please keep going until the user’s query is completely resolved, before ending your turn and yielding back to the user. Your thinking should be thorou…

It is. But not because Cursor is dumb, because how it handles the context. I can tell you from CC side that longer - and wider - your context, Claude is dumber (which makes sense). There are other threads where Cursor folks are trying to figure out what’s happening with contexts (what is being passe…

i disagree. I know when it is using chatgpt as i found a hack based on its last information knowledge date. Chatgpt can still function but there is one AI that comes in and just drags its feet, acts like its stoned and clearly does not want to be at work that day.

no its not a skill issue, it’s happened to me intermittently. full agent mode asking questions. I was thinking that when the usage is high they might get it to ask a lot of questions to slow you down. just a thought.

@Jonneal3 thank you for the report. This post seems to have slipped through due to many posts of last 24h. Could you share if your original request included either: Question on why an issue occurs? or Request to fix an issue? (This would be the best approach as asking a question about an issue l…

The auto mode has controls for which models to use. Its in the settings. There is definitely a model that is lazy. (I hate that guy). Consider sometimes my ill form prompts generater fully operational programs. To other times im begging it stop asking me questions or confirming every step.

Maybe. Can’t you force it to tell you which model it is using? Usually its Sonnet 3.7 or 4 for me, but maybe its lying. When it does use gpt its obvious, but it also tells me now, so I am not surprised.

Insane Laziness In Agent

Ideas Feedback

MidnightOak July 22, 2025, 3:26pm 21

Where is this setting?

Topic		Replies	Views
GPT 5 is really bad (at least in Cursor) Discussions	142	15985	January 13, 2026
Opus Max (500usd) No Fix vs GPT o3 Fixed in 3 msgs Discussions	30	1335	July 26, 2025
GPT-5 paralyzed by excessive thinking? Discussions	8	809	September 17, 2025
Model auto agent mode does never follow the prompt and lies Discussions auto-mode	10	228	March 18, 2026
People, Your Honest Opinion Discussions	23	2836	March 18, 2025

Insane Laziness In Agent

Related topics