careful

Destructive shell commands like rm -rf and git reset --hard are intercepted by a pre-use Bash hook that scans input for dangerous patterns and returns a permission decision before execution. This mechanism stops accidental data loss across development, staging, and production environments.

Yes, per-command overrides are available for destructive commands in production environments. When the safety hook flags a risky operation, you can explicitly choose to override the warning and proceed with the execution.

This production safety hook covers shell workflows including rm -rf, git reset --hard, kubectl delete, and SQL DROP statements. It applies across development, staging, and production environments to guard risky operations.

You set up guardrails for risky shell operations by enabling safety mode, which activates a pre-tool hook to flag dangerous commands. The hook surfaces a clear warning prompt requiring explicit confirmation before running commands like git reset or kubectl delete.

Yes, this risk-management tool logs incidents when it blocks destructive commands. After the pre-use Bash hook scans input and returns a permission decision, it records the event to maintain an audit trail of flagged operations.