careful

Detect destructive bash commands and return permission warnings before execution.

11|2|Updated Mar 31, 2026
One-click install
npx skills add https://github.com/ajsai47/holyclaude --skill careful-ajsai47
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: careful
Source: https://github.com/ajsai47/holyclaude/tree/main/skills/team/careful
Command: npx skills add https://github.com/ajsai47/holyclaude --skill careful-ajsai47

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires bash, grep, sed, python3.

What problem does it solve?

Accidental execution of destructive commands like recursive file deletions, database drops, forced git pushes, or Kubernetes resource deletions can cause irreversible data loss, uncommitted work deletion, or production outages when working in shared environments, production systems, or during live debugging sessions.

Core Features & Use Cases

  • Multi-Category Destructive Command Detection: Identifies 8+ high-risk command patterns including recursive file deletes, SQL DROP and TRUNCATE statements, git force pushes and hard resets, kubectl delete operations, and Docker force-remove and prune commands.
  • Smart Safe Exceptions: Automatically allows common build artifact deletions such as node_modules, dist, .next, and pycache without triggering false positive warnings.
  • User Override Support: All warnings can be overridden if you intentionally need to run a destructive command, with no forced blocking of intentional actions.
  • Use Case: When debugging a live production Kubernetes cluster, if you accidentally type kubectl delete pod instead of kubectl get pod, the skill will immediately warn you of the destructive action before it executes, preventing accidental service downtime.

Quick Start

Activate the careful skill before running any commands in your production environment or shared codebase to receive real-time warnings for destructive operations.

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?

Preventing accidental data loss from destructive bash commands involves applying guardrails that detect high-risk patterns like recursive file deletions and SQL DROP statements, returning descriptive warning messages before execution. This skill monitors 8 categories of destructive operations.

Does this safety guardrail allow safe deletions like removing node_modules?

This safety guardrail allows safe deletions like removing node_modules, dist, .next, and __pycache__ directories without triggering false positive warnings, while still blocking risky recursive file deletions and uncommitted work deletion.

Can I override the warning if I intentionally need to run a destructive git force push?

You can override the warning if you intentionally need to run a destructive git force push or hard reset, as all warnings can be overridden without forced blocking of intentional actions, ensuring your workflow remains uninterrupted.

What types of Kubernetes and Docker destructive commands does it detect?

It detects Kubernetes destructive commands like kubectl delete operations and Docker force-remove and prune commands, preventing accidental service downtime and irreversible data loss during live production debugging sessions.

Do I need python3 and bash installed to use this command safety checker?

You need python3 and bash installed to use this command safety checker, as these are required dependencies for executing the pattern matching logic that scans your commands for destructive file system, SQL, git, and Kubernetes operations.