guard

Emit warnings before destructive commands and enforce directory edit boundaries.

Updated Apr 15, 2026
One-click install
npx skills add https://github.com/256javy/safia --skill guard-256javy
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: guard
Source: https://github.com/256javy/safia/tree/main/.claude/skills/gstack/guard
Command: npx skills add https://github.com/256javy/safia --skill guard-256javy

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Guards against dangerous edits by emitting real-time warnings when destructive commands are issued and by enforcing a boundary around editable directories.

Core Features & Use Cases

  • Destructive command warnings: Intercepts risky commands and prompts for confirmation, reducing accidental system damage.
  • Directory edit boundary (freeze): Restricts edits to a user-specified directory to prevent unintended changes elsewhere.
  • Cross-skill safety integration: Works with the sibling careful and freeze skills to provide holistic safety in production debugging and maintenance tasks.

Quick Start

Ask the user which directory to protect and enable guard mode with destructive-command warnings and a directory edit boundary.

Frequently Asked Questions about guard

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

FAQPage Schema
How do I prevent accidental destructive bash commands from executing in production?

Enforcing directory edit boundaries restricts file modifications to a user-specified directory, preventing unintended changes elsewhere and keeping edits confined within a defined path.

Can I restrict bash edits to a specific directory to protect sensitive paths?

Enforcing directory edit boundaries restricts file modifications to a user-specified directory, preventing unintended changes elsewhere and keeping edits confined within a defined path.

How do I enable guard mode to warn before destructive commands in my workflow?

It uses PreToolUse hooks to intercept commands in real-time, applying specifically to workflows that touch production systems and sensitive directories to ensure safe execution.

Does this safety mechanism work for production system debugging and maintenance?

It uses PreToolUse hooks to intercept commands in real-time, applying specifically to workflows that touch production systems and sensitive directories to ensure safe execution.

What is the best way to coordinate workflow security when running risky shell commands?

The best way to coordinate workflow security is integrating cross-skill safety boundaries, working with sibling careful and freeze skills to provide holistic protection during sensitive tasks.