guard

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

2|Updated Apr 4, 2026
One-click install
npx skills add https://github.com/adityasingh2400/SwarmSell --skill guard-adityasingh2400
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: guard
Source: https://github.com/adityasingh2400/SwarmSell/tree/main/.claude/skills/guard
Command: npx skills add https://github.com/adityasingh2400/SwarmSell --skill guard-adityasingh2400

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Guard mode reduces risk when working on production or debugging live systems by enforcing destructive command warnings and a directory boundary to prevent unauthorized edits.

Core Features & Use Cases

  • Destructive-command warnings: prompts before executing dangerous commands like rm -rf, DROP TABLE, and forced pushes.
  • Directory-scoped edits: blocks edits outside a user-specified path, preventing accidental changes to critical areas.
  • Compatibility: combines the functionalities of the sibling /careful and /freeze skills to provide an integrated safety layer during maintenance or debugging tasks.
  • Use Case: IT engineers safely perform live-debugging while ensuring no unwanted changes spill outside the authorized directory.

Quick Start

Ask the user for the directory to restrict edits to, then activate guard mode to enable both protections.

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 commands during live debugging on production?

You can restrict directory edits during maintenance by specifying a target path before activating guard mode, which blocks any unauthorized changes from spilling outside that boundary.

How do I restrict directory edits during maintenance to prevent changes outside a specific path?

You can restrict directory edits during maintenance by specifying a target path before activating guard mode, which blocks any unauthorized changes from spilling outside that boundary.

Does guard mode support cross-skill dependencies with other workflow security tools?

Guard mode protects production environments by combining destructive-command warnings with directory-bound edits, intercepting dangerous commands and blocking unauthorized changes outside a specified path.

What types of destructive commands does guard mode intercept?

Guard mode intercepts destructive commands like rm -rf, DROP TABLE, and forced pushes, prompting the user for confirmation before allowing them to execute in the production environment.

Can I establish a robust safety workflow without needing external dependencies?

You can establish a robust safety workflow without external dependencies because guard mode integrates pre-tool hooks, user-question prompts, and directory boundaries natively.