careful

Detect destructive Bash and operational commands before execution.

1|Updated May 5, 2026

Search Tags:#git #production #bash #safety #kubernetes #guardrails #destructive command

Ready to run

npx skills add https://github.com/yashs33244/my-mac-claude --skill careful-yashs33244

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: careful
Source: https://github.com/yashs33244/my-mac-claude/tree/main/skills/gstack/careful
Command: npx skills add https://github.com/yashs33244/my-mac-claude --skill careful-yashs33244

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill prevents costly mistakes by warning you before you run destructive shell commands that can delete data, remove infrastructure resources, or rewrite history.

Core Features & Use Cases

Destructive command detection: Flags high-risk operations like rm -rf/-r, SQL DROP/TRUNCATE, force-push, git reset --hard, and common kubectl/docker deletes and prunes.
Safety guardrails with overrides: If a destructive pattern is detected, it asks for confirmation so you can cancel or proceed intentionally.
Production and shared-environment support: Helps when you’re “being careful,” “safety mode,” or working in environments where mistakes impact others.

Quick Start

Ask to run in safety mode by telling the assistant: “be careful and warn me before any destructive commands like rm -rf or git reset --hard.”

Frequently Asked Questions about careful

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

FAQPage Schema

How do I prevent destructive bash commands from running in production?▼

To prevent destructive bash commands in production, you can use a safety guardrail that detects high-risk operations like rm -rf, git reset --hard, or SQL DROP/TRUNCATE and asks for confirmation before execution. This pattern-based matching enforces a PreToolUse permission decision to warn you.

What destructive commands should I guard against during infrastructure maintenance?▼

During infrastructure maintenance, you should guard against rm -rf, SQL DROP/TRUNCATE, git force-push, git reset --hard, and common kubectl or docker deletes and prunes. Detecting these destructive commands prevents unintended data loss and history rewrites.

Can I get a warning before running git force-push or kubectl delete in shared environments?▼

Yes, you can get a warning before running git force-push or kubectl delete in shared environments by applying a safety mode guardrail. It detects these destructive operational commands and requires a confirmation step so you can cancel or proceed intentionally.

How do I enable safety mode for shell-based production debugging workflows?▼

To enable safety mode for shell-based production debugging workflows, prompt the assistant to be careful and warn before destructive commands. This activates pattern-based matching with predefined exceptions to intercept high-risk bash commands before execution.

Are there exceptions to the destructive command detection patterns?▼

Yes, the destructive command detection applies pattern-based matching with predefined exceptions. This means certain safe commands matching similar patterns are allowed through, while genuinely dangerous operations like rm -rf or SQL DROP still trigger a confirmation warning.