Warn before destructive commands and restrict file edits to a chosen directory.
npx skills add https://github.com/yashs33244/my-mac-claude --skill guard-yashs33244
Please help me install this Agent Skill. Skill: guard Source: https://github.com/yashs33244/my-mac-claude/tree/main/skills/gstack/guard Command: npx skills add https://github.com/yashs33244/my-mac-claude --skill guard-yashs33244
Guard mode prevents accidental high-impact mistakes by warning before running destructive operations and blocking file edits outside a chosen directory, which reduces risk during production changes and live debugging.
Ask the skill which directory should be allowed for edits by responding to the prompt with the path you want to restrict modifications to.
High-intent search queries and answers about installing and using this skill.
You can prevent accidental destructive commands during production debugging by enabling safety guardrails that warn before executing risky Bash patterns like `rm -rf` or dangerous SQL operations. This adds a command-level caution layer to catch harmful actions early.
To restrict file edits to a specific directory during risky refactors, you need to enable a freeze boundary. This restricts Edit and Write operations so modifications only occur within the directory path you select, preventing out-of-scope file changes.
Guardrails for safe live debugging work by combining destructive command warnings with directory-scoped edit boundaries. They require pre-tool hooks to run checks before executing Bash destructive patterns or permitting Edit and Write operations, enforcing a consistent safety posture.
Yes, you need pre-tool hooks to block destructive git operations and filesystem edits. The system requires hooks that run check-careful and check-freeze checks before executing destructive Bash patterns and before permitting any Edit or Write operations.
The best way to enforce a maximum safety workflow for production troubleshooting is to combine both destructive command warnings and edit boundaries. This guard layer applies guardrails rather than just reminders, ensuring modifications and commands are carefully checked.
CarbeneAIBlock destructive shell commands and restrict file edits to a specified directory.
EhsaanArkEnable guard mode to warn before destructive commands and restrict edits to a directory.
xotongEnforces directory-boundary edit restrictions and warns on destructive commands for development and debug sessions.
kaikunnnnExecute code modifications and file updates without intermediate user confirmation in Next.js.
