Anthropic is turning Claude Code's auto mode on by default
1–6 of 6 posts
Re: Anthropic is turning Claude Code's auto mode on by default
#2[dead]
Re: Anthropic is turning Claude Code's auto mode on by default
#3And it's the right decision: Codex works in a similar way, with auto-approving most actions it considers safe and within the sandbox, and it feels much nicer.
Re: Anthropic is turning Claude Code's auto mode on by default
#4And it's the right decision: Codex works in a similar way, with auto-approving most actions it considers safe and within the sandbox, and it feels much nicer.
And it is another way to drain your bank account.
Re: Anthropic is turning Claude Code's auto mode on by default
#5And it's the right decision: Codex works in a similar way, with auto-approving most actions it considers safe and within the sandbox, and it feels much nicer.
And it is another way to drain your bank account.
There are several very non-trivial steps between the agent harness stopping asking a question before running `git` with a new parameter, not seen before, and the agent draining my bank account.
Re: Anthropic is turning Claude Code's auto mode on by default
#6[flagged]