guard

Warn on destructive commands and restrict edits to a specified directory.

1|Updated Feb 27, 2026
One-click install
npx skills add https://github.com/geoffreywoo/clawfable --skill guard-geoffreywoo
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: guard
Source: https://github.com/geoffreywoo/clawfable/tree/main/.claude/skills/gstack/guard
Command: npx skills add https://github.com/geoffreywoo/clawfable --skill guard-geoffreywoo

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Guard mode provides a unified safety layer that surfaces destructive command warnings and enforces directory-bound edit restrictions, reducing the risk of inadvertent production changes.

Core Features & Use Cases

  • Destructive command warnings: alerts before executing dangerous commands like rm -rf or force-push.
  • Directory edit boundary: prevents edits outside a user-specified path, protecting critical files.
  • Unified safety workflow: combines careful and freeze capabilities into a single, repeatable setup for devops and SREs.

Quick Start

Run the gstack setup and specify the directory to restrict edits.

Frequently Asked Questions about guard

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

FAQPage Schema
How do I prevent accidental edits outside a specific directory in production?

To prevent accidental edits outside a production directory, enforce directory edit boundaries by specifying a restricted path during setup, which resolves absolute paths and blocks external file modifications to protect critical files.

Can I get warnings before executing destructive bash commands like rm -rf?

Yes, you can receive warnings before executing destructive bash commands by hooking into bash-based safety checks, which alert you before dangerous actions like rm -rf or force-push are executed.

What is a unified safety workflow for devops and SREs?

A unified safety workflow for devops and SREs combines destructive command warnings with directory edit restrictions into a single, repeatable setup, reducing the risk of inadvertent production changes.

How do I set up a restricted directory boundary for system protection?

To set up a restricted directory boundary, run the setup workflow to specify the target directory, resolve its absolute path, and store the boundary for ongoing bash-based safety checks.

Does guard mode work without additional dependencies or components?

Yes, guard mode works without additional dependencies or components, implementing its destructive command warnings and edit boundary enforcement entirely through built-in bash hooks.