careful

Warn and require confirmation before destructive shell commands execute.

1|Updated Mar 27, 2026
One-click install
npx skills add https://github.com/gebl/anvil-skill-marketplace --skill careful-gebl
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: careful
Source: https://github.com/gebl/anvil-skill-marketplace/tree/main/generated/claude/skills/careful
Command: npx skills add https://github.com/gebl/anvil-skill-marketplace --skill careful-gebl

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill provides safety guardrails for destructive commands, warning users before executing operations like rm -rf, DROP TABLE, force-push, git reset --hard, kubectl delete, and similar destructive operations. Users can override warnings when necessary.

Core Features & Use Cases

  • Warns before destructive commands to prevent accidental data loss.
  • Allows safe override to proceed when the user is sure.
  • Integrates with Bash workflows to safeguard development and production tasks.

Quick Start

Type a destructive command like rm -rf /var/log to see a warning and confirm before execution.

Frequently Asked Questions about careful

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I prevent accidental data loss from destructive bash commands like rm -rf?

To prevent accidental data loss from destructive bash commands like rm -rf, you need a pre-execution analysis hook that detects risky patterns and issues a warning requiring confirmation before proceeding.

Can I add a warning prompt before running git reset --hard or kubectl delete?

Yes, you can add a warning prompt before running git reset --hard or kubectl delete by using a pre-execution hook that identifies risky command patterns and requests explicit user confirmation.

How do I override a destructive command warning if I am sure I want to proceed?

You can override a destructive command warning when you are certain of the action, as the safety guardrail allows users to bypass the prompt and proceed with executing the risky operation.

Does this safety hook work across both development and production kubectl workflows?

Yes, this safety hook works across both development and production kubectl workflows, applying its pre-execution analysis to detect and warn against destructive operations in any environment.

What is the best way to safeguard shell workflows against force-push and DROP TABLE operations?

The best way to safeguard shell workflows against force-push and DROP TABLE operations is to implement a permission decision hook that intercepts risky patterns and pauses execution for user validation.