guard

You can prevent destructive commands during production debugging by enforcing safety checks that warn before executing dangerous operations like rm -rf, DROP TABLE, or force-push, and by restricting edits to a user-specified directory.

To restrict file modifications to a specific directory during system maintenance, the Skill enforces an edit boundary by prompting for a target directory and storing this freeze boundary in a state file, blocking unauthorized edits outside that path.

An edit boundary protects live systems by restricting all file modifications to a user-specified directory, preventing risky changes outside that defined path. It is applied during maintenance and code reviews to block unauthorized edits.

Yes, you can enforce both destructive-command warnings and directory restrictions simultaneously. The Skill provides a combined safety workflow that integrates both protections together, offering end-to-end guardrails during maintenance and debugging tasks.

Yes, you need to specify a target directory when prompted to activate guard mode. Once provided, both destructive-command warnings and edit boundary enforcement are applied to your workflow.