careful

Warn before destructive Bash, Git, Kubernetes, and Docker commands with explicit override.

Updated Mar 26, 2026
One-click install
npx skills add https://github.com/mrkhachaturov/rkstack --skill careful-mrkhachaturov
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: careful
Source: https://github.com/mrkhachaturov/rkstack/tree/main/skills/careful
Command: npx skills add https://github.com/mrkhachaturov/rkstack --skill careful-mrkhachaturov

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Safety guardrails alert before destructive commands (rm -rf, git reset --hard, kubectl delete, and similar operations) and provide an explicit override mechanism, helping teams avoid accidental data loss in live environments.

Core Features & Use Cases

  • PreToolUse hook integration that scans commands for risky patterns and prompts for confirmation when necessary.
  • Destructive-command detection across common tools (bash, git, kubernetes, docker) to reduce human error in critical tasks.
  • Override workflow with explicit user consent to proceed in sanctioned scenarios, preserving safety without halting essential work.

Quick Start

Enable careful mode to receive warnings before destructive commands and override prompts as needed.

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?

Preventing accidental data loss from destructive bash commands like rm -rf requires safety guardrails that scan for risky patterns and prompt for explicit confirmation before execution. This skill uses a PreToolUse hook to intercept such operations.

Does this safety guardrail work with Kubernetes and Docker workflows?

Yes, this safety guardrail works with Kubernetes and Docker workflows. It detects destructive actions like kubectl delete across bash, git, kubernetes, and docker environments, prompting for explicit override before executing commands that risk data loss.

Can I override the destructive command warning when I need to proceed?

Yes, you can override the destructive command warning when you need to proceed. The skill provides an explicit user consent workflow for sanctioned scenarios, preserving safety guardrails without halting essential work in live environments.

How does a PreToolUse hook detect risky git commands like git reset --hard?

A PreToolUse hook detects risky git commands like git reset --hard by scanning executed commands for destructive patterns. It enforces pre-checks and environment detection, warning you before the operation proceeds to protect critical systems.

What is the best way to add workflow safety guardrails for live environments?

The best way to add workflow safety guardrails for live environments is integrating a PreToolUse hook that enforces pre-checks and safe-execution defaults. This proactively warns before destructive operations and requires explicit override consent.