careful

Warn before destructive Bash commands via a PreToolUse hook.

Updated Feb 5, 2026
One-click install
npx skills add https://github.com/CaltexBevo/ihe-pulse --skill careful-caltexbevo
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: careful
Source: https://github.com/CaltexBevo/ihe-pulse/tree/main/.claude/skills/gstack.bak/careful
Command: npx skills add https://github.com/CaltexBevo/ihe-pulse --skill careful-caltexbevo

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Destructive operations in shells and automation can cause irreversible damage. This Skill provides real-time guardrails that warn before executing dangerous commands such as rm -rf, git reset --hard, kubectl delete, or similar actions, reducing the chance of accidental data loss.

Core Features & Use Cases

  • Warns before destructive commands and prompts for confirmation to cancel or proceed.
  • Allows per-command overrides when you explicitly intend to perform a destructive operation in a controlled environment.
  • Integrates with a PreToolUse workflow by inspecting Bash commands through a dedicated check script to help keep prod and shared systems safe.

Quick Start

Enable the careful skill in your workflow and issue Bash commands with the hook that warns before destructive actions and prompts you to confirm or cancel.

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, you need real-time guardrails that warn before executing dangerous operations. This Skill intercepts commands like rm -rf, prompts for confirmation, and provides safe defaults to cancel or proceed.

Can I get a warning before running destructive git reset or kubectl delete operations?

Yes, you can get a warning before running destructive git reset or kubectl delete operations. The Skill detects dangerous patterns through a PreToolUse hook and prompts you to confirm or cancel before execution proceeds.

Does this destructive command warning system work in both interactive shells and automation?

This destructive command warning system works in both interactive shells and automation. It inspects Bash commands through a dedicated check script, returning a permission decision to help keep prod and shared systems safe during dangerous operations.

How do I override the warning when I intentionally need to run a destructive command?

You can override the warning when you intentionally need to run a destructive command by using the per-command override feature. This allows you to explicitly proceed with dangerous operations in a controlled environment after the prompt appears.

What is the best way to add safety guardrails to dangerous docker prune commands?

The best way to add safety guardrails to dangerous docker prune commands is to use a PreToolUse hook that detects destructive patterns and prompts for confirmation. This approach intercepts the command before execution, reducing the chance of accidental data loss.

When should I not use a destructive command warning hook in my workflow?

You should not use a destructive command warning hook when you require fully unattended automation without any prompts, as it pauses execution to ask for confirmation. It is designed for interactive shells and controlled environments where user override is possible.