The Jev-enator

What it does

Add three Claude Code hooks that flag dangerous commands, failed tool calls, and incomplete work with Jev.

How you can use it

A developer runs the setup tool to add these checks. The tool copies your old settings first so you can restore them later. It requires an access key generated by the TypeSafe AI service. This key connects the software to the service. It is not a personal login password.

The software sends three questions in a single request to evaluate the work. The service chooses one answer to allow, warn, or block the action. If the top choices are too close, it names no problem. The software blocks actions marked dangerous. However, the AI cannot catch every bad command.