safety-guard

Block destructive commands and restrict file edits to specified directories.

Updated Nov 19, 2025
One-click install
npx skills add https://github.com/Sake-Team/SmartSake --skill safety-guard-sake-team
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: safety-guard
Source: https://github.com/Sake-Team/SmartSake/tree/main/backup/skills/safety-guard
Command: npx skills add https://github.com/Sake-Team/SmartSake --skill safety-guard-sake-team

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill prevents accidental destructive actions on production systems, providing an extra layer of safety during automated or autonomous agent operations.

Core Features & Use Cases

  • Careful Mode: Blocks destructive commands and warns before execution.
  • Freeze Mode: Restricts file edits to a specific directory tree.
  • Guard Mode: Combines both Careful and Freeze modes for maximum protection.
  • Unlock Command: Safely turn off safety measures when necessary.
  • Integration: Compatible with 'codex -a never' sessions and ECC 2.0 observability risk scoring.

Quick Start

Use the /safety-guard guard --dir src/api/ --allow-read-all command to ensure your agent only writes to the 'src/api/' directory.

Frequently Asked Questions about safety-guard

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

FAQPage Schema
How do I prevent destructive commands in production environments during automated operations?

To prevent destructive commands in production environments, use Careful Mode to block dangerous execution and warn against risky actions. This adds an extra layer of safety during autonomous agent operations.

How do I restrict automated agent file edits to a specific directory tree?

You can restrict file edits to a specific directory tree by using Freeze Mode or Guard Mode. Executing a command like /safety-guard guard --dir src/api/ ensures your agent only writes to that specified directory.

Can I manually override safety guardrails if an automated process gets blocked?

Yes, you can manually override safety guardrails if an automated process gets blocked. The Skill provides an Unlock Command that safely turns off safety measures when necessary for manual control.

Does production protection work with ECC 2.0 observability risk scoring?

Yes, production protection works with ECC 2.0 observability risk scoring. The Skill is fully compatible with ECC 2.0 risk scoring and 'codex -a never' sessions to enhance compliance.

What is the best way to combine command blocking and directory restriction for system security?

The best way to combine command blocking and directory restriction for system security is Guard Mode. It merges both Careful and Freeze modes to provide maximum protection against accidental destructive actions.