[Bug] Excessive safety filtering flags legitimate code refactoring requests
Bug Description
What the f is wrong with you guys? Since one month you're really fing it up and you're pi$$ing Max subscribers like me off. What's so difficult to build a proper model and not CONSTANTLY flag proper prompts as violations? It's getting tiring, I already stopped trying your Fable model because of it as even a simple refactoring of code that you wrote yourself with Opus was being blocked. Now you're starting the same shit job in Opus models? Do you care about your customer? Just say the word if you want me to cancel my subscription and move to OpenAI or some Chinese model which costs me a fraction of yours but doesn't bother me all the time with a completely shitty job of identifying when a normal human user is just trying to fix bugs and develop new features for my SaaS app!
Environment Info
- Platform: darwin
- Terminal: iTerm.app
- Version: 2.1.198
- Feedback ID: 8c423929-d513-41ed-ad99-42194b48f304
Errors
[]Why is the "endogamy-shadow-flag-c306" still on a worktree/branch? I thought we finished coding it in another terminal window?
⏺ API Error: Claude Code is unable to respond to this request, which appears to violate our Usage Policy (https://www.anthropic.com/legal/aup). Please double press esc to edit your last message or start a new session for Claude Code to assist with a different task.
Request ID: req_011Ccp7RBQL6B8qNKDWC9boK
✻ Brewed for 6s
❯ Why is the "endogamy-shadow-flag-c306" still on a worktree/branch? I thought we finished coding it already?
⏺ API Error: Claude Code is unable to respond to this request, which appears to violate our Usage Policy (https://www.anthropic.com/legal/aup). Please double press esc to edit your last message or start a new session for Claude Code to assist with a different task.
Request ID: req_011Ccp7VEXTfQwHyYDXoFix9