careful

Warn before executing destructive shell commands via a PreToolUse hook.

Updated Apr 22, 2026
One-click install
npx skills add https://github.com/diiviikk5/rezops --skill careful-diiviikk5
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: careful
Source: https://github.com/diiviikk5/rezops/tree/main/careful
Command: npx skills add https://github.com/diiviikk5/rezops --skill careful-diiviikk5

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires python3.

What problem does it solve?

Safety guardrails to prevent destructive shell commands from causing data loss or service disruptions. It warns before executing dangerous commands and allows user override when appropriate.

Core Features & Use Cases

  • Pre-emptive warnings for destructive operations like rm -rf, git reset --hard, kubectl delete, and similar commands.
  • Optional override to proceed after a warning, with session-scoped guards.
  • Lightweight hook-based protection that can be integrated into tool execution pipelines.

Quick Start

Activate careful mode to warn before destructive commands during tool runs.

Frequently Asked Questions about careful

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

FAQPage Schema
How do I warn before executing destructive shell commands like rm -rf?

To warn before executing destructive shell commands like rm -rf, you need a safety mode hook that detects dangerous patterns and prompts for confirmation. This prevents data loss by intercepting operations before they run.

Can I get a warning before running kubectl delete or git reset --hard in a shared environment?

Yes, you can get a warning before running kubectl delete or git reset --hard in a shared environment by using a PreToolUse hook. This guardrail detects destructive commands and allows an optional override to proceed safely.

Do I need python3 to set up guardrails for destructive bash commands?

Yes, you need python3 installed to set up guardrails for destructive bash commands. The hook uses python3 to detect destructive patterns and log incidents locally before executing shell operations.

What is the best way to prevent data loss from accidental docker prune commands?

The best way to prevent data loss from accidental docker prune commands is a lightweight hook-based protection system. It pre-empts destructive operations, allows safe exceptions for certain artifacts, and logs incidents locally.

How does a safety guardrail handle safe exceptions for destructive commands?

Safety guardrails handle safe exceptions by allowing user overrides for certain artifacts while maintaining session-scoped guards. The system detects destructive patterns, warns the user, and logs the incident locally before proceeding.

When should I not use automated warnings for shell operations?

You should not use automated warnings for shell operations when executing fully trusted, automated scripts where manual override prompts would disrupt pipelines. The session-scoped guards require user interaction to confirm destructive commands.