careful

A pre-execution warning for destructive commands works by using a PreToolUse Bash hook to analyze commands before they run. It detects operations like git reset --hard and kubectl delete, then pauses to require explicit user confirmation before proceeding.

Yes, you can override the safety warning to proceed with destructive commands. The Skill provides optional override support that allows you to bypass the pre-execution prompt and execute operations like git reset --hard after giving explicit confirmation.

Yes, this destructive command guard works for kubectl delete operations across prod and shared environments. It is applicable to both interactive shells and automation pipelines, ensuring warnings trigger before risky kubectl actions are executed.

To log usage analytics for overridden destructive commands in devops, the Skill automatically records usage data when the pre-execution hook triggers. This provides governance and auditing records for every warning event and subsequent override action.

The best way to prevent accidental data loss from rm -rf in an automation pipeline is to implement a PreToolUse Bash hook. This mechanism intercepts destructive commands, enforces warning prompts, and logs usage analytics for auditing without stopping valid pipeline flows.

You should not use a pre-execution warning hook for bash safety in fully autonomous, non-interactive pipelines where pausing for a prompt would cause a timeout. While overrides are supported, requiring confirmation can block automated flows if not properly configured.