Most days I switch between Claude Code and Codex. The annoying part: the same MCP servers set up twice, the same tokens in two files, and no real limit on what either agent can touch.
So I built Fenzi. One sandbox framework, one MCP list, whichever agent I open.
https://t.co/qpb1SWEbr3
Claude Code just added a setting that says no for you when nobody is there to answer a permission prompt.
When you walk away, should your agent stop and wait, or say no and keep going?
@Im_IrushiK These reset limits are tricky because they make you work 10x more on things that are not necessarily useful. Better waste my usage than maxing out for no reason lol
Claude Code's "don't ask again" can save the whole command you approved, password included. Researchers found 1,122 of these files on public GitHub with secrets inside.
Do you still click "don't ask again"?
@lydiahallie@lydiahallie it would be even better if from this /storage command, we could have Claude go through its past work to see any temporary files that were created but left there or unused packages that were downloaded
Postman's desktop app quietly installed its own plugins into Claude Code and Codex during an update.
Would you notice if another app added something to your coding agent?