safety-guard

Monitors Bash, Write, Edit and MultiEdit tool calls to restrict destructive commands and unintended changes.

1|Updated Apr 6, 2026
One-click install
npx skills add https://github.com/zero3041/PREP --skill safety-guard-zero3041
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: safety-guard
Source: https://github.com/zero3041/PREP/tree/main/.claude/skills/skills/safety-guard
Command: npx skills add https://github.com/zero3041/PREP --skill safety-guard-zero3041

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill helps in safeguarding against accidental or destructive actions when working with production systems or autonomous agents, thereby preserving system integrity.

Core Features & Use Cases

  • Prevent Destructive Commands: Blocks harmful operations like rm -rf, git push --force, and git reset --hard.
  • Restrictive File Editing: Limits write/edit access to specific directories to prevent unintended changes.
  • Comprehensive Protection: Offers three modes (Careful, Freeze, Guard) for varying levels of protection.
  • Integration and Logging: Defaults for codex -a never sessions, logs blocked actions, and integrates with ECC 2.0 for observability risk scoring.

Quick Start

Activate the safety-guard skill with the command /safety-guard guard --dir src/api/ --allow-read-all.

Frequently Asked Questions about safety-guard

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

FAQPage Schema
How do I prevent autonomous agents from running destructive bash commands in production?

You can prevent destructive bash commands by monitoring and restricting execution of operations like rm -rf and git push --force within Bash, Write, and Edit tool calls. This safeguards against unauthorized or unintended system changes during autonomous agent operations.

Can I restrict file editing access to specific directories only?

Yes, you can restrict file editing access to specific directories. The skill limits write and edit access to prevent unintended changes, configurable via command flags such as specifying a target directory during activation.

What is the best way to monitor and block git reset --hard operations in development environments?

The best way to block git reset --hard operations is using a system protection mode that monitors tool calls for destructive commands. It offers Careful, Freeze, and Guard modes to provide varying levels of operational restriction.

Does operation monitoring integrate with risk scoring systems for observability?

Yes, operation monitoring integrates with risk scoring systems for enhanced observability. It defaults to codex -a never sessions, logs blocked actions, and integrates with ECC 2.0 to provide observability risk scoring.

When should I use Freeze mode versus Guard mode for environment safeguarding?

Use Freeze mode or Guard mode based on your required level of environment safeguarding. The skill offers three modes—Careful, Freeze, and Guard—allowing you to apply precise operational restrictions appropriate for your development or production context.