Warn before executing destructive Bash commands via a PreToolUse hook.
npx skills add https://github.com/cloudofgeorge/AI-hands --skill careful-cloudofgeorge
Please help me install this Agent Skill. Skill: careful Source: https://github.com/cloudofgeorge/AI-hands/tree/main/skills/gstack/careful Command: npx skills add https://github.com/cloudofgeorge/AI-hands --skill careful-cloudofgeorge
Guardrails are essential to prevent accidental or destructive commands from wreaking havoc on systems. The careful skill provides real-time warnings to prevent risky actions in Bash, enabling safer command execution in shared or production environments.
Enable careful mode to automatically warn before destructive Bash commands in your session.
High-intent search queries and answers about installing and using this skill.
You can prevent accidental system damage from destructive Bash commands by enabling a mode that uses a PreToolUse hook to scan operations before execution. This guardrails approach warns you of risky actions, supports an override flow, and logs usage events for audits in shared environments.
A PreToolUse hook works by intercepting Bash commands and running a check script before execution. If destructive patterns are detected, it pauses the operation to warn you, offering a choice to proceed with an override or cancel the action entirely to ensure safety.
Yes, you can override the warning when running deployment scripts. When the system detects a destructive command, it presents a warning and offers an override flow, allowing you to explicitly choose to proceed with the risky operation or cancel it to maintain safety.
Yes, the destructive command warning system supports audit logging for shared environments. It logs usage metrics to a local analytics directory, tracking when warnings are triggered and overridden, which helps improve safety protocols and provides an audit trail for production systems.
The best way to add safety guardrails to interactive Bash sessions is to enable a warning mode that automatically scans for destructive patterns. This prevents accidental damage by intercepting risky commands before execution and requiring explicit user confirmation to proceed.
tropeksWarn before executing destructive shell commands via a PreToolUse hook.
GTC6244Inspect shell commands for destructive patterns before execution.
HurleySkEnforce layered safety checks on Claude Code write and edit operations.
gwpjpDetect and block destructive commands, malicious packages, and secret exposure during agent operations.
