Cursor blocks basic prompts with usage guidelines error
Cursor blocks nearly every prompt with a 'usage guidelines' error, preventing developers from completing simple coding edits. Users report the guardrails trigger on non-sensitive, normal requests, making the AI coding assistant unusable for routine tasks. No clear fix or way to bypass the overzealous blocking has been provided.
Evidence
“A couple messages in with the new bot, he became unresponsive. Messages to him failed to sync with my other devices and on reopening the app they show “failed to send””
“We're having trouble connecting to the model provider. This might be temporary - please try again in a moment.”
“Same problem as the past few previous days, all my bots again stopped talking... talk to a bot... they talk back”
“getting the code is often easier, than understanding whether it is actually the solution.”
“It did not stop that task when it hit the on-demand cap I set. It did not return an 'exceeds credit limit' stop message.”
“Cursor Ask/Agent is stuck indefinitely on “Reconnecting…” and does not return any response. ... Network Diagnostics now passes all checks, including Chat and Agent, but actual Ask/Agent requests remain stuck on “Reconnecting…”.”
“Tasks that it used to handle quite well are now often handled in a really stupid way — losing context, making obvious mistakes, going in the wrong direction, or simply failing to finish the task.”
“Fast mode is selected anyway”
“The individual tool response was shorter, but on average, the task used more tokens and took longer. We saved tokens locally and spent more globally.”
“Continuously getting the same response which is 'Grok Bot Failed to Respond'”
“I’m getting the error “An unexpected error occurred” when trying to use chat/agents in Cursor. Request ID: 6dffb4c9-40ab-46fd-b704-65bc0c6fb0e2 I suspect this might be related to archived projects/chats”
“if it is possible to use opencode go api in cursor without having a cursor subscription? plz guide if possible”
“We are unable to complete this request because it was blocked under the model provider's usage guidelines. Try a less sensitive prompt.”
“an agent doing a 'cleanup' nobody asked for and wiping a whole profile”
“sometimes developers end up with code they can't fully explain or maintain”
“frustrated by two patterns in current AI coding environments: cloud vendor lock-in that wipes project memory when switching providers, and platforms adding heavy markups on top of user API keys.”
“they are biasing the auto mode to choose grok for all tasks no matter easy or hard”
“This work flow really took away my passion and joy of writing and building software systems.”
“struggling to understand what I’m really getting from Cursor Start. Would upgrading to Cursor Pro actually add much value in my case? I already have Claude Max 5x and Codex Pro”
“I've added my API key under Settings > Models > Google , but the editor seems locked into using Cursor's built-in free models anyway.”
“Why would cursor block this request I am just asking to very basic coding update request.”
“It seems the Fast mode for Composer and Grok went slower than before? Anyone experience the same?”
“the app isnt connecting as well as not finding agent”
“Timed out waiting for the agent host bridge to register; please retry.”
“my token usage has exploded. What am I doing wrong?”
“I faced a problem of screenshotting everything and explaining the context and theri relations to claude code again and again like for little UI chanegs and such or any bug.”
“about 80% of the dev candidates that I interview tell me that they aren't writing much code themselves anymore - they are directing agents instead.”
“every bot fails when I try to send a message”
“Continuously showing 'Taking Longer than Expected'”
“What I have to do today for Cloud Agents is messy - rules are in User Rules, I initially synced skills from ~/.cursor/skills, but they can go stale... so I started putting skills and custom agents in a git repo, and install them on recurring Cloud Env builds. It works, but it's a workaround, not a neat, clean solution.”
“For the last several updates it's consistently done this annoying thing where it will add a curly bracket that needs to be removed before the code can compile.”
“Cursor Grok 4.6 (High) model is using 3rd party API usage instead of First Party Model usage. I have reached 100% of both usage even though i have barely used any 3rd party models.”
“has anyone else noticed agents explicitly refusing to perform tasks?”
“the 10 days of limit reach has no explanation. I use it the same way i always use it.”
“I'm in auto and you decided it's best for my API usage to go from 0% to 30% with one call with Auto....what the......just.....damn it man.”
“It won't stop describing statistical adjustments as 'fudge' (I get the term but it uses it in non sensible ways and obscuring what it days with the word fudge)”
“I am trying to do a task but it showing taking longer than expected”
“I don't even call them pull requests. I call them prompt requests.”