I have everything I can think of added the Allow List Terminal Commands, in agent settings, and auto execution is set to always proceed. No matter what I still have to repetitively click allow over and over. simple commands pass, but commands with | ; and “ almost always ask for approval. it makes it impossible to run multiple agents at a time…
3 Likes
Can you confirm if this is a bug or is intentional? I’ve spent a lot of time trying to figure out how to let the system automate, but it seems like the settings imported during startup override my user settings. I’ve sent a bug report but not heard back.
Hello @Ben_Hardin, welcome to AI Forum.
Apart from the steps you mentioned, you could try the following:
- Go to the Allow List configuration in settings and add a single asterisk * on a new line. Some users report this forces the parser to treat everything as allowed, overriding the step that flags complex commands as dangerous. Warning: This bypasses almost all safety checks.
- Sometimes the global settings desync from the workspace settings. Click your Profile Icon (top right) Quick Settings and ensure Always Proceed is selected there as well.
- After changing the settings, fully restart the agent (or the IDE) to flush the cached execution policy.