careful

You can prevent accidental execution of destructive shell commands by using a PreToolUse hook that identifies risky patterns like rm -rf, git reset --hard, and kubectl delete before they run. It emits a permission decision prompting you to confirm or override the operation.

Adding safety guardrails for production environments in bash involves detecting destructive operations before execution and emitting a warning. This acts as a safety mechanism that prompts you to selectively override the warning when you are certain the operation is safe to proceed.

Yes, you can override command warnings when you know a destructive git push -f is safe. When the system detects a risky pattern, it prompts you with a warning and allows you to selectively override it to proceed with the execution.

Yes, destructive command detection works for kubectl delete operations. The mechanism identifies risky patterns associated with cluster and tool changes, including kubectl delete, raising a confirmation prompt to guard the command before it actually runs.

The best way to protect shared environments from accidental data loss is implementing context-aware protection that intercepts risky file deletions and database modifications before they run. This prevents destructive shell commands from executing accidentally by requiring explicit confirmation.