careful

To add a confirmation prompt before destructive shell commands, use a pre-tool hook that inspects the command and emits a permissionDecision ask signal. This guardrail intercepts dangerous patterns like rm -rf and prompts for explicit override before execution.

This safety guardrail intercepts destructive command patterns such as rm -rf, git reset --hard, kubectl delete, and docker system prune. A pre-tool hook analyzes the input command to detect these actions and prevent accidental data loss or downtime.

Yes, you can use command guardrails to prevent accidental data loss in automation pipelines. The hook-based safety mechanism analyzes commands and returns a permission decision to prevent unsafe runs automatically, ensuring destructive actions require explicit approval.

To override a blocked destructive command, you explicitly approve or cancel the action when the warning message appears. The override workflow presents a clear warning to prevent accidental damage, allowing you to confirm and continue execution if intentional.

Yes, this command warning system works across different interactive shells and automation pipelines. It integrates a pre-tool hook that analyzes the input command regardless of the shell environment to consistently intercept and warn before destructive actions.