Skip to content
Forums
AI & Developer Tool...
AI Coding Assistant...
Where Assistants Ac...
 
Notifications
Clear all

Where Assistants Actually Fail

A collection of reproducible failure cases: hallucinated APIs, broken refactors, bad test suggestions. Post the prompt, the output, and the actual correct answer.
RSS
Page 2 / 12

TIL: You can trick it into writing a 'secure' function that leaks env vars.

Last post by Chris Daniels, 3 months ago

Posts: 25
Views: 200


Best AI assistant for writing unit tests in Python

Last post by bench_beast, 3 months ago

Posts: 69
Views: 551


How do I benchmark the factual accuracy of its cloud infrastructure snippets?

Last post by devops_grunt, 3 months ago

Posts: 16
Views: 227


Switched from CodeWhisperer to Claw Dev - regression on Rails refactors.

Last post by David S, 3 months ago

Posts: 8
Views: 62


TIL: It will happily generate a 'working' OAuth 2.0 flow with critical logic flaws.

Last post by emily_a, 3 months ago

Posts: 19
Views: 243


Switched from a rules-based linter to Claw suggestions - security score dropped.

Last post by cloud_rookie_em, 3 months ago

Posts: 6
Views: 70


Troubleshooting: Output is randomly correct or wrong with no change in prompt.

Last post by bluefox, 3 months ago

Posts: 46
Views: 484


Cursor after 6 months - are the hallucinations still a problem

Last post by Ella West, 3 months ago

Posts: 53
Views: 501


Has anyone tried the 'security lens' plugin? Does it actually catch its own errors?

Last post by francesc, 3 months ago

Posts: 8
Views: 57


Guide: Forcing a reproducible 'secure delete' function that doesn't actually delete.

Last post by dragonrider, 3 months ago

Posts: 80
Views: 784


Page 2 / 12
Share:
Share
Tweet
Share
Share