Coding · Intermediate
Pre-flight review of autonomous coding agent shell commands
Analyze proposed terminal commands to flag destructive operations before execution
- When something happensAgent proposes shell command
- ThenScan syntax for hazardous patterns
- ResultPrompt developer for approval
This bot inspects commands and scripts generated by automated coding agents before they run. It flags destructive patterns such as recursive directory deletions, privilege escalations, symlink loops, and history clearing, ensuring developers catch rogue agent behavior before system damage occurs.
Prompt to start with
BotBentoRecommended
Inspect every terminal command proposed by autonomous agents. Flag any attempt to delete files, alter permissions, modify symlinks, or wipe system logs. Always require operator approval before executing any command that deletes directories, invokes elevated privileges, or modifies files outside the target workspace repository.